Target detection method and device for mower, electronic equipment and storage medium
Through the lidar sensor, point cloud data acquisition and projection processing is performed on the surface of the lawn mower shell, and combined with the classification model to conduct inclement weather detection, solving the problem of high design difficulty and excessive volume in the lawn mower target detection method, achieving more efficient weather recognition and equipment simplification.
Patent Information
- Application Number
- CN202411110647.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Priority Date
- 2023-10-31
- Filing Date
- 2024-08-13
- Publication Date
- 2024-11-26
AI Technical Summary
The existing target detection methods of lawn mowers have problems such as high difficulty in designing equipment and excessive equipment volume. Especially when it is necessary to detect bad weather in outdoor environments, adding sensors will increase design complexity and cost.
By configuring a lidar sensor to detect its shell surface, obtain point cloud data, and target recognition through projection processing and classification models, the detection and classification of bad weather is achieved, avoiding the need to add additional sensors.
It reduces the difficulty and volume of equipment design, while improving the universality and accuracy of target detection, solving the problems of excessive equipment size and design complexity.
Smart Images

Figure CN119027649A_ABST
Abstract
Description
Technical Field
[0001] The embodiments of the present application relate to the computer field, and in particular, to a target detection method and device for a lawn mower, an electronic device, and a storage medium. Background Art
[0002] At present, for some devices that need to be used outdoors or in other unobstructed scenarios, in order to avoid adverse effects on device operation due to environmental factors, additional sensors can be added to detect these environmental targets. For example, a rain sensor can be used to detect rain, or a visible light sensor can be added to train the model for discrimination.
[0003] However, the above-mentioned method of performing target detection by adding redundant sensors requires setting the added sensors at specific locations based on the characteristics of the environmental targets to be detected, which not only increases the design difficulty and engineering costs, but also causes the device to be too large, affecting the user experience of the device.
[0004] It can be seen that the target detection method in the related art has the problems of high difficulty in equipment design and excessively large equipment size. Summary of the invention
[0005] The embodiments of the present application provide a target detection method and device, an electronic device and a storage medium for a lawn mower, so as to at least solve the problems of high difficulty in equipment design and excessive size of the target detection method in the related art.
[0006] According to one aspect of an embodiment of the present application, a target detection method for a lawn mower is provided, comprising: obtaining a target frame set obtained by detecting a sensor shell surface of the target sensor through a target sensor, wherein each target frame in the target frame set is a single-frame point cloud obtained by a detection by the target sensor, and the target sensor is a sensor configured on the lawn mower for generating a point cloud; performing a projection operation on the target frame set to obtain a target projection image, wherein the projection operation is an operation of projecting onto a specified plane; inputting the target projection image into a trained target classification model to obtain a target classification result, wherein the target classification model is a classification model for distinguishing multiple preset categories, and the target classification result is used to indicate a preset category among the multiple preset categories that matches the target frame set.
[0007] According to another aspect of an embodiment of the present application, a target detection device for a lawn mower is also provided, comprising: an acquisition unit, for acquiring a target frame set obtained by detecting a sensor shell surface of the target sensor through a target sensor, wherein each target frame in the target frame set is a single-frame point cloud obtained by a detection by the target sensor, and the target sensor is a sensor configured on the lawn mower for generating a point cloud; a first execution unit, for performing a projection operation on the target frame set to obtain a target projection image, wherein the projection operation is an operation of projecting onto a specified plane; an input unit, for inputting the target projection image into a trained target classification model to obtain a target classification result, wherein the target classification model is a classification model for distinguishing multiple preset categories, and the target classification result is used to indicate a preset category among the multiple preset categories that matches the target frame set.
[0008] According to another aspect of an embodiment of the present application, a lawn mower control method is also provided, wherein the lawn mower includes a laser radar, and the method includes: after detecting bad weather through the laser radar, the lawn mower returns to a base station or continues to work.
[0009] According to another aspect of an embodiment of the present application, a lawn mower control method is also provided, wherein the lawn mower includes a laser radar, and the method includes: generating and sending weather information after detecting bad weather through the laser radar; and the lawn mower returns to the base station or continues to work after receiving the control signal.
[0010] According to another aspect of an embodiment of the present application, a lawn mower control method is also provided, which is applied to a base station. The method includes: after the base station receives weather information, sending a control signal to control the lawn mower to return to the base station or continue working.
[0011] According to another aspect of an embodiment of the present application, a lawn mower control method is also provided, which is applied to a user end, and the method includes: after the user end receives weather information, sending a control signal to control the lawn mower to return to the base station or continue working.
[0012] According to another aspect of an embodiment of the present application, a lawn mower control method is also provided, wherein the lawn mower includes a laser radar, and the method includes: in a case where there are water stains on the shell surface of the laser radar, obtaining a water signal obtained by the laser radar detecting the shell surface at least once, and judging the current weather based on the obtained water signal; and controlling the lawn mower to perform corresponding actions according to the current weather; the method of controlling the lawn mower to perform corresponding actions according to the current weather includes: in a case where it is judged that the current weather is bad weather, performing a shielding action on at least part of the lawn mower.
[0013] According to another aspect of an embodiment of the present application, a lawn mower control method is also provided, wherein the lawn mower includes a laser radar, and the method includes: in a case where there are water stains on the shell surface of the laser radar, obtaining water signals obtained by the laser radar detecting the shell surface at least twice within a set total time period, and determining the current weather based on the obtained water signals; controlling the lawn mower to perform corresponding actions according to the current weather; the controlling the lawn mower to perform corresponding actions according to the current weather includes: in a case where it is determined that the current weather is bad weather, performing a shielding action on at least part of the lawn mower.
[0014] According to another aspect of an embodiment of the present application, a lawn mower control method is also provided, wherein the lawn mower includes a laser radar, and the method includes: when there are water stains on the shell surface of the laser radar, obtaining a water signal obtained by the laser radar detecting the shell surface at least once; when the coverage value of a single water signal increased on the shell surface per unit time is greater than a preset value, controlling the lawn mower to continue mowing according to a preset cutting path and cutting mode.
[0015] According to another aspect of an embodiment of the present application, a computer-readable storage medium is provided, in which a computer program is stored, wherein the computer program is configured to execute the above-mentioned target detection method for a lawn mower or the above-mentioned lawn mower control method when running.
[0016] According to another aspect of an embodiment of the present application, there is also provided an electronic device, including a memory and a processor, wherein the memory stores a computer program, and the processor is configured to execute the target detection method for a lawn mower or the lawn mower control method through the computer program.
[0017] Through the embodiment of the present application, a target frame set is obtained by detecting the surface of a sensor shell of a target sensor through a configured sensor, wherein each target frame in the target frame set is a single frame point cloud obtained by a detection by the target sensor, and the target sensor is a sensor configured on the lawn mower for generating a point cloud; a projection operation is performed on the target frame set to obtain a target projection image, wherein the projection operation is an operation of projecting onto a specified plane; the target projection image is input into a trained target classification model to obtain a target classification result, wherein the target classification model is a classification model for distinguishing multiple preset categories, and the target classification result is used to indicate multiple preset categories. The preset category that matches the target frame set can adapt to the detection of one or more environmental targets, improve the universality of target detection, and have low requirements on the setting position of the sensor, which can reduce the difficulty of equipment design. At the same time, the sensor used can be a configured sensor for obstacle detection, without the need to add additional sensors, which can achieve the technical effect of reducing the difficulty of equipment design and reducing the size of the equipment, thereby solving the problems of high equipment design difficulty and excessive equipment size in the target detection method in the related art. BRIEF DESCRIPTION OF THE DRAWINGS
[0018] Figure 1 is a hardware structure block diagram of an optional lawn mower according to an embodiment of the present application;
[0019] Figure 2 is a flow chart of an optional target detection method for a lawn mower according to an embodiment of the present application;
[0020] Figure 3 is a schematic diagram of an optional target detection method for a lawn mower according to an embodiment of the present application;
[0021] Figure 4 is a schematic diagram of another optional target detection method for a lawn mower according to an embodiment of the present application;
[0022] Figure 5 is a flow chart of another optional target detection method for a lawn mower according to an embodiment of the present application;
[0023] Figure 6 is a flow chart of another optional target detection method for a lawn mower according to an embodiment of the present application;
[0024] Figure 7 is a structural block diagram of a target detection device for a lawn mower according to an embodiment of the present application;
[0025] Figure 8 is a structural block diagram of a computer system of an optional electronic device according to an embodiment of the present application;
[0026] Fig. 9 is a flow chart of an optional lawn mower control method according to an embodiment of the present application;
[0027] Fig.10 is a flow chart of another optional lawn mower control method according to an embodiment of the present application;
[0028] Fig.11 It is a flowchart of another optional lawn mower control method according to an embodiment of the present application. DETAILED DESCRIPTION
[0029] The embodiments of the present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.
[0030] It should be noted that the terms "first", "second", etc. in the description and claims of the embodiments of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence.
[0031] The method embodiments provided in the embodiments of the present application can be executed in a lawn mower or a similar processing device. Taking running on a lawn mower as an example, Figure 1 1 is a hardware structure block diagram of an optional lawn mower according to an embodiment of the present application. Figure 1 As shown, the lawn mower may include one or more ( Figure 1 In addition, for some smart lawn mowers, they may also include: a sensor 108, a processor 110 (the processor 110 may include but is not limited to a processing device such as a microprocessor MCU or a programmable logic device FPGA) and a memory 112 for storing data, wherein the above-mentioned lawn mower may also include a transmission device and an input and output device for communication functions. It can be understood by those skilled in the art that Figure 1 The structure shown is only for illustration and does not limit the structure of the above lawn mower. Figure 1 More or fewer components as shown, or with Figure 1 Different configurations are shown.
[0032] The memory 112 can be used to store computer programs, for example, software programs and modules of application software, such as the computer program corresponding to the target detection method for a lawn mower in the embodiment of the present application. The processor 110 executes various functional applications and data processing by running the computer program stored in the memory 112, that is, the target detection method for a lawn mower described above is realized. The memory 112 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some examples, the memory 112 may further include a memory remotely arranged relative to the processor 110, and these remote memories may be connected to the mobile terminal via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.
[0033] The transmission device is used to receive or send data via a network. The specific example of the above network may include a wireless network provided by a communication provider of the lawn mower. In one example, the transmission device includes a NIC (Network Interface Controller), which can be connected to other network devices through a base station so as to communicate with the Internet. In one example, the transmission device can be an RF (Radio Frequency) module, which is used to communicate with the Internet wirelessly.
[0034] According to one aspect of an embodiment of the present application, a target detection method for a lawn mower is provided. The method can be executed by a server or a lawn mower alone, or by a server and a lawn mower together, or by a terminal device that has a direct or indirect communication connection with the lawn mower. Taking the lawn mower as an example to execute the target detection method for a lawn mower in this embodiment, Figure 2 is a flow chart of an optional target detection method for a lawn mower according to an embodiment of the present application, such as Figure 2 As shown, the process includes the following steps:
[0035] Step S202 : acquiring a target frame set obtained by detecting the sensor shell surface of the target sensor through the target sensor.
[0036] The target detection method for a lawn mower in this embodiment can be applied to a scenario where a lawn mower or similar equipment is detected. The detected target can be an environmental target, and the environmental target can be a target in an outdoor environment that has an impact on the operation of the equipment, such as rain, dust, grass clippings or other targets, or a target that has an impact on the operation of the equipment in some special environments (inside a production workshop). Although the lawn mower is used as an example for explanation in this embodiment, the target detection method in this embodiment is also used for other similar equipment (represented by target equipment). In some examples of this embodiment, rain detection of a lawn mower (rain detection, for scenes with rain, the equipment will be exposed to rain) is used as an example for explanation.
[0037] Here, the lawn mower is also called a weeder, a lawn mower, a lawn trimmer, etc. A lawn mower is a mechanical tool used to mow lawns, vegetation, etc. It may include a cutter, an engine, a running wheel, a running mechanism, a blade, an armrest, a control part, etc. It may be an intelligent lawn mower, that is, a lawn mower that does not require manual control by a user, or a lawn mower that is manually controlled by an intelligent assistant. In this embodiment, there is no limitation on the components included in the lawn mower, as long as the mowing function and the target detection function can be realized.
[0038] In order to avoid the impact of adding redundant rain sensors on the difficulty of equipment design, equipment volume, etc., in this embodiment, the point cloud set obtained by detecting the surface of the sensor shell based on the sensor installed on the lawn mower can be projected, and the classification model can be used to perform target recognition on the projected image obtained by the projection. This can adapt to the detection of one or more environmental targets, improve the universality of target detection, and have lower requirements on the setting location of the sensor, which can reduce the difficulty of equipment design. At the same time, the sensor used can be a configured sensor for obstacle detection, without the need to add additional sensors, which can reduce the difficulty of equipment design and reduce the size of the equipment.
[0039] For a lawn mower, a target sensor may be provided on the lawn mower, and the target sensor may be a sensor already equipped on the lawn mower (may be a configured sensor for generating a point cloud), for example, a sensor for obstacle detection, and the sensor may be a laser radar sensor (referred to as laser radar) or other types of sensors. The lawn mower (or a processor thereon) may obtain a target frame set obtained by detecting the surface of the sensor shell through the target sensor, where each target frame in the target frame set is a single frame point cloud obtained by a detection by the target sensor.
[0040] Here, LiDAR is a device that measures distance and creates a three-dimensional point cloud map by sending a laser beam and receiving reflected light. In this device, a laser scanner sends a laser beam and records the time and position of the reflected light to generate point cloud data. After the laser beam is reflected back to the device, the device measures the time and determines the position where the laser beam was sent, thereby calculating the distance traveled by the reflected laser beam. By scanning the laser beam to obtain reflected light in all directions in space, the device can obtain a point cloud in three-dimensional space.
[0041] For example, Figure 3 As shown, laser radar can include three major systems: laser emission, laser reception, and information processing. These three systems complement each other to form a closed sensing loop. Taking the modeling of objects through laser as an example, first, the excitation source in the laser emission system periodically drives the laser to emit laser pulses. The laser modulator controls the direction and number of lines of the emitted laser through the beam controller. Finally, the laser is emitted to the target object through the emission optical system; the scanning system is responsible for rotating at a stable speed to scan the plane and generate real-time plane map information; the photoelectric detector in the laser receiving system receives the laser reflected back by the target object and generates a receiving signal; the received signal in the information processing system is amplified and converted to digital and analog, and is calculated by the information processing module to obtain the surface morphology, physical properties and other characteristics of the target, and finally establish an object model.
[0042] In this embodiment, the target frame set may include a point cloud set obtained by the target sensor performing multiple detections on the surface of the sensor shell within a certain period of time. The detection period of the target sensor detecting the surface of the sensor shell may be pre-configured. The multiple detections may be continuous or non-continuous. The target frame set is not limited in this embodiment.
[0043] Optionally, the target sensor can be an integrated sensing device. In addition to point cloud detection, the target sensor can also be used for at least one of the following: navigation, positioning, obstacle avoidance, and the operations applied to navigation, positioning, obstacle avoidance, etc. can be performed based on the detected point cloud, but the processing logic and judgment method are different. In addition, the target sensor is also applicable to other operations that can be performed based on the detected point cloud.
[0044] Optionally, in order to facilitate target detection, the surface of the sensor shell can be a plane and / or a curved surface. For different detection targets, the type of sensor shell surface can be flexibly selected to ensure the accuracy of target detection while avoiding the impact of target accumulation on the performance or service life of the equipment. The target sensor can be set at an angle to the horizontal plane at the front end area of the top of the lawn mower. Here, the target sensor is set at an angle to the horizontal plane, which means that the detection center line of the target sensor is at a certain angle to the horizontal plane (that is, the target sensor has a certain inclination angle, for example, the inclination angle is 30 degrees), for example, 30 degrees, 45 degrees or other angles. The target sensor is set at an angle to the horizontal plane at the front end area of the top of the lawn mower, and the detected area is in an upper position to avoid inaccurate detection results due to occlusion and other reasons (for example, curved surface).
[0045] For example, a lawn mower may be equipped with a laser radar, which has a protective shell, and the surface of the protective shell is flat or spherical, and the laser radar is installed at a certain tilt angle. Accumulate the laser radar point cloud, accumulating n frames in total (n is a positive integer greater than or equal to 2):
[0046] PointCloud sum ={frame1,frame2,…,frame i}
[0047] frame i ={point1,point2,…,point j}
[0048] Among them, PointCloud sum It can be represented as a LiDAR accumulated point cloud (i.e., a target frame set); frame i It can be expressed as a single-frame point cloud of the laser radar, 0≤i≤n; point j It can be represented as a lidar point, 0≤j≤m.
[0049] Step S204: performing a projection operation on the target frame set to obtain a target projection image.
[0050] In order to improve the convenience of target detection based on the target frame set, a projection operation can be first performed on the target frame set, that is, each position point in the target frame set is projected onto a specified plane, and the target frame set is converted into a two-dimensional image, so that the target detection method can be used for target detection. The specified plane can be a horizontal plane or other planes. Here, for the projection method of the target frame set, the position point information of each position point in the target frame set can be used as the pixel value of the corresponding projection position (pixel position). If the projection positions of multiple position points are the same, the position point information of multiple position points with the same projection position can be fused. The projection method of the target frame set is not limited in this embodiment.
[0051] Optionally, the projection operation on the target frame set may be performed by directly converting the collected point cloud data into a projection image in a specified plane through the function of the target sensor, and then fusing the projection images in multiple conversions to obtain the target projection image. Alternatively, the single-frame point cloud collected multiple times may be accumulated to obtain the target frame set, and then the projection operation may be performed on the target frame set as a whole. Other projection operation execution methods may also be used, which are not limited in this embodiment.
[0052] Step S206, inputting the target projection image into the trained target classification model to obtain a target classification result.
[0053] After obtaining the target projection image, the target projection image can be input into the trained target classification model to obtain the target classification result. The target classification model can be a classification model for distinguishing multiple preset categories, and the target classification result can be used to indicate the preset category that matches the target frame set among the multiple preset categories, that is, the preset category that matches the target projection image.
[0054] The target classification model can be a convolutional neural network model (CNN) or other types of classification models. Here, a convolutional neural network is a type of feedforward neural network that includes convolution calculations and has a deep structure. It is one of the representative algorithms of deep learning. This type of method can effectively utilize the mature network structure in two-dimensional images and use massive two-dimensional image annotation data for pre-training, thereby achieving good classification results.
[0055] Optionally, multiple preset categories can be set according to the environmental targets to be detected. For example, for a scene where only rain detection is performed, the multiple preset categories can include two types: rain and non-rain. For rain, it can be divided into multiple levels, for example, the first level (corresponding to light rain), the second level (corresponding to moderate rain), the third level (corresponding to heavy rain), or other division methods, and each level can correspond to a preset category. In addition to rain, the detected environmental targets can also include dust, grass clippings, mud, etc. The preset categories can include types, levels, etc. corresponding to these environmental targets, as long as the environmental targets to be detected and the levels of the environmental targets can be distinguished by the image features of the projected image.
[0056] Optionally, in order to improve the diversity of target detection, preset types can be configured for a variety of different environmental targets, and multiple preset categories are used to represent at least one of the following targets: rain, hail, snow, sandstorm, fog, etc. In addition, one or more levels can be set for each environmental target to indicate the intensity of the corresponding environmental target.
[0057] Through the above steps, a target frame set is obtained by detecting the sensor shell surface of the target sensor through the target sensor, wherein each target frame in the target frame set is a single frame point cloud obtained by a detection by the target sensor, and the target sensor is a sensor configured on the lawn mower for generating a point cloud; a projection operation is performed on the target frame set to obtain a target projection image, wherein the projection operation is an operation of projecting onto a specified plane; the target projection image is input into a trained target classification model to obtain a target classification result, wherein the target classification model is a classification model for distinguishing multiple preset categories, and the target classification result is used to indicate a preset category among multiple preset categories that matches the target frame set, thereby solving the problems of large equipment volume and high equipment design difficulty in the target detection method in the related art, reducing the difficulty of equipment design, and reducing the equipment volume.
[0058] In an exemplary embodiment, acquiring a target frame set obtained by detecting a sensor housing surface of a target sensor through a target sensor includes:
[0059] S11, acquiring multiple groups of echo signals obtained by transmitting detection signals multiple times to the surface of the sensor shell through the target sensor, wherein each group of echo signals in the multiple groups of echo signals includes two echo signals corresponding to a detection signal transmitted once;
[0060] S12, fusing two echo signals in each group of echo signals into a corresponding single-frame point cloud to obtain a target frame set.
[0061] In this embodiment, when the surface of the sensor housing is detected by the target sensor, multiple groups of echo signals can be obtained by transmitting multiple detection signals to the surface of the sensor housing through the target sensor. The time interval for transmitting the detection signals can be fixed or not completely the same. For each detection signal transmitted, a group of echo signals can be obtained, and each group of echo signals includes two echo signals corresponding to the detection signal transmitted once. In this embodiment, in order to ensure the rationality of target detection, one echo signal in each group of echo signals is an echo signal reflected by the surface of the sensor housing, and the other echo signal is an echo signal reflected by the detected target.
[0062] For example, according to the model of the laser radar, the original two echo data of the laser radar are used for processing, the first echo is the echo of the protective shell, and the second echo is the echo of the environmental target on the protective shell.
[0063] For each group of echo signals obtained, two of the echo signals can be fused into a corresponding single-frame point cloud to obtain a target frame set. The fusion method can be to calculate the difference between the two, or to fuse the two in combination with other information. The fusion method of the echo signals is not limited in this embodiment.
[0064] Through this embodiment, two echo signals obtained by transmitting a detection signal once are fused to obtain a single-frame point cloud in a target frame set, which can improve the convenience and accuracy of point cloud data acquisition.
[0065] In an exemplary embodiment, two echo signals in each group of echo signals are fused into a corresponding single-frame point cloud to obtain a target frame set, including:
[0066] S21, based on the signal difference at the same position point of two echo signals in each group of echo signals, determine the single frame point cloud corresponding to each group of echo signals to obtain a target frame set, wherein, in the target frame set, the value of a position point corresponds to the signal difference between the two echo signal position points corresponding to the position point.
[0067] When fusing two echo signals in each group of echo signals, a single-frame point cloud corresponding to each group of echo signals can be determined based on the signal difference at the same position point of the two echo signals in each group of echo signals, that is, a single-frame point cloud obtained by each detection, thereby obtaining a target frame set. Here, determining the single-frame point cloud corresponding to each group of echo signals based on the signal difference at the same position point of the two echo signals in each group of echo signals can be: according to a preset conversion relationship, the signal difference at the same position point of the two echo signals in each group of echo signals is converted into the value of the same position point in the single-frame point cloud corresponding to each group of echo signals.
[0068] Optionally, in the target frame set, the value of a position point can be used to identify the degree of dirtiness of the position point (which can be considered as a dirtiness tag of the position point). If the degree of dirtiness of a position point is higher, the signal value of the echo signal corresponding to the position point is larger. On the contrary, if the degree of dirtiness of a position point is lower, the signal value of the echo signal corresponding to the position point is smaller. The above-mentioned preset conversion relationship can be set based on this feature. Here, the preset conversion relationship can be: a positive correlation relationship or other conversion relationships, as long as the degree of dirtiness of each position point can be accurately characterized.
[0069] For example, the dirty mark on the surface of the lidar shell can be obtained based on multiple lidar echoes. By combining two radar echoes, the dirty mark tag of the radar (used to indicate the degree of obstruction of the corresponding position) can be obtained. For the n frames accumulated by the lidar point cloud, any point in each frame j It can be expressed as:
[0070]
[0071] Among them, ρ is the modulus length, θ is the polar angle, is the azimuth, tag is the dirty label, which is used to identify the degree of occlusion of the corresponding position, 1≤j≤m.
[0072] Through this embodiment, by determining the single-frame point cloud corresponding to each group of echo signals based on the signal difference at the same position point of two echo signals in each group of echo signals, the convenience of point cloud data determination can be improved, and the ability of point cloud data to represent environmental targets can be improved.
[0073] In an exemplary embodiment, acquiring a target frame set obtained by detecting a sensor housing surface of a target sensor through a target sensor includes:
[0074] S31, obtaining a target frame set obtained by detecting a laser radar shell surface of a multi-dimensional laser radar through a multi-dimensional laser radar, wherein the target sensor is a multi-dimensional laser radar.
[0075] In this embodiment, the target sensor may be a multi-dimensional laser radar, so that the characteristics of the detection target can be more accurately characterized based on the acquired multi-dimensional information. Here, the sensor shell surface of the target sensor is the laser radar shell surface of the multi-dimensional laser radar. Correspondingly, the method of acquiring the target frame set may be: acquiring the target frame set obtained by detecting the laser radar shell surface through the multi-dimensional laser radar. Optionally, the multi-dimensional laser radar is a three-dimensional (Three Dimensions) laser radar.
[0076] Through this embodiment, the use of a multi-dimensional laser radar to acquire a cumulative point cloud can improve the ability of the cumulative point cloud to represent the detection target, thereby improving the accuracy of target detection.
[0077] In an exemplary embodiment, each position point in each target frame is represented by a set of position parameters, such as the aforementioned modulus, polar angle and azimuth, or x-coordinate, y-coordinate and z-coordinate; the value of each position point in each target frame is used to represent the degree of occlusion of the corresponding position point, which can be considered as a dirty label of the corresponding position point.
[0078] Correspondingly, before performing the projection operation on the target frame set, the method further includes:
[0079] S41, performing a filtering operation on a target frame set based on at least one position parameter to obtain a filtered target frame set, wherein the at least one position parameter belongs to a group of position parameters.
[0080] Taking into account the installation angle and gap of the target sensor (for example, the radar installation angle and gap), dirt may be hidden on the target sensor, the surface of the sensor shell or other objects within the detectable range of the target sensor. In this regard, it is possible to consider filtering the target frame set based on all or part of the position parameters in a set of position parameters (that is, at least one position parameter). This can improve the accuracy of target detection while reducing the amount of data to be processed. Here, at least one position parameter belongs to a set of position parameters.
[0081] In this embodiment, a filtering operation can be performed on the target frame set based on at least one position parameter to obtain a filtered target frame set. Here, there can be multiple position parameters used to perform the filtering operation on the target frame set. The multiple position parameters can be used to perform the filtering operation on the target frame set in sequence. After the target frame set is filtered using the last position parameter of the multiple position parameters, the filtered target frame set can be obtained. Optionally, in order to improve the processing efficiency, the multiple position parameters can be used to perform the filtering operation on the target frame set respectively, and the intersection of the filtered target frame sets is determined as the filtered target frame set.
[0082] For example, Figure 4 As shown, two position parameters can be used to perform filtering operations on the target frame set. In the case of performing filtering operations on the target frame set in sequence, the filtering operation performed using the latter position parameter is for: the target frame set obtained after the filtering operation is performed on the target frame set using the previous position parameter (including unfiltered collaboration). In the case of performing filtering operations on the target frame set separately, after performing filtering operations on the target frame set using each position parameter separately, two filtered target frame sets can be obtained, and the intersection of the two filtered target frame sets is determined as the final filtered target frame set.
[0083] Through this embodiment, at least one position parameter is used to perform a transfer operation on a position point in the accumulated point cloud, which can improve the accuracy of target detection while reducing the amount of data required to be processed.
[0084] In an exemplary embodiment, performing a filtering operation on a target frame set based on at least one position parameter to obtain a filtered target frame set includes at least one of the following:
[0085] S51, filtering out the position points in the target frame set whose corresponding modulus length is not a specified value, and obtaining the target frame set after modulus length filtering;
[0086] S52, filtering out the position points in the target frame set whose corresponding polar angles are greater than or equal to the specified angle, to obtain the target frame set after polar angle filtering.
[0087] In this embodiment, considering that the installation angle and gap of the target sensor may hide dirt, the position points in the target frame set may be filtered according to at least one of the module length and the polar angle. When performing the filtering operation according to the module length, the position points in the target frame set whose corresponding module length is not a specified value may be filtered out to obtain the target frame set after the module length filtering, wherein the specified value may be set according to the position of the coordinate origin of the spherical coordinate system, for example, the specified value is the value of the module length corresponding to the surface of the sensor shell. When performing the filtering operation according to the polar angle, the position points in the target frame set whose corresponding polar angle is greater than or equal to the specified angle may be filtered out to obtain the target frame set after the polar angle filtering, wherein the specified angle is determined based on the setting position of the sensor shell, and the covered area may be the range of the gap that does not include the edge of the sensor shell.
[0088] For example, in order to reduce the influence of points not on the radome, the accumulated point cloud can be filtered based on the modulus ρ, and the filtered points can be the positions where the modulus is not 0 in the accumulated point cloud:
[0089]
[0090] Among them, 1 means retaining; 0 means discarding.
[0091] For example, in order to reduce the impact of dirty gaps, the accumulated point cloud can be filtered based on the polar angle θ, and the filtered point cloud can be the point cloud with a polar angle exceeding the set threshold θ. thres Location point:
[0092]
[0093] Among them, 1 means keep; 0 means discard; θ thres Indicates the angle of the dirty gap relative to the z-axis.
[0094] Through this embodiment, the position points in the accumulated point cloud are filtered based on at least one of the modulus length and the polar angle, which can reduce the influence of points on the sensor shell and the influence of dirty gaps, and improve the convenience of position point filtering.
[0095] In an exemplary embodiment, a set of position parameters includes three axis coordinates in a rectangular coordinate system, and the position points in the target frame set acquired by the target sensor can be represented by a spherical coordinate system. In this case, the position points in the spherical coordinate system can be converted to position points in the rectangular coordinate system through coordinate transformation. The tilt angle of the target sensor includes a roll angle (roll direction) and a pitch angle (pitch direction).
[0096] In order to reduce the influence of ground points on target detection, the position points in the target frame set are filtered using the tilt angle of the target sensor. Correspondingly, a filtering operation is performed on the target frame set based on at least one position parameter to obtain a filtered target frame set, including:
[0097] S61, using a first rotation matrix and a second rotation matrix to transform the coordinate values of the three axis coordinates of each position point in the target frame set to obtain a target frame set after coordinate transformation, wherein the first rotation matrix is a rotation matrix determined according to the roll angle, and the second rotation matrix is a rotation matrix determined according to the pitch angle;
[0098] S62, filtering out the position points whose corresponding z-axis coordinates are less than a preset value in the target frame set after coordinate transformation, to obtain the target frame set after coordinate filtering.
[0099] In this embodiment, it is considered that the coordinate values of the three-axis coordinates of the position points in the target frame set are the coordinate values in the sensor coordinate system. For this, it is necessary to convert the coordinate values in the sensor coordinate system into the coordinate values in the world coordinate system before performing point cloud filtering. The rotation matrix used to convert the coordinate values in the sensor coordinate system of the target sensor into the coordinate values in the world coordinate system includes a first rotation matrix and a second rotation matrix, wherein the first rotation matrix is a rotation matrix determined according to the roll angle (i.e., a rotation matrix in the roll direction), and the second rotation matrix is a rotation matrix determined according to the pitch angle (i.e., a rotation matrix in the pitch direction).
[0100] When performing a filtering operation on a target frame set, the first rotation matrix and the second rotation matrix can be first used to transform the coordinates of each axis of each position point in each target frame to obtain a target frame set after coordinate transformation. At this time, each position point in the target frame set is represented by the coordinate values of the three coordinate axes in the world coordinate system; then, in the target frame set after coordinate transformation, the position points whose corresponding z-axis coordinates are less than a preset value are filtered out to obtain a target frame set after coordinate filtering, where the preset value is a value greater than or equal to the z-axis coordinate corresponding to the ground.
[0101] For example, in order to reduce the influence of ground points, the radar tilt angle (roll direction and pitch direction) is used for filtering: the point cloud after coordinate transformation is transformed using the rotation matrix in the roll direction and the pitch direction, and filtering is performed based on the transformed point cloud, where the rotation matrix in the roll direction and the pitch direction are shown in formula (1) and formula (2):
[0102]
[0103] Among them, R X and R Y They can be expressed as rotation matrices in the roll and pitch directions, respectively. The formula for coordinate transformation using the rotation matrices in the roll and pitch directions is shown in formula (3):
[0104]
[0105] The method for filtering the point cloud after coordinate transformation is:
[0106]
[0107] Among them, 1 means to keep, 0 means to discard; rotation Represents the coordinates after rotation.
[0108] Through this embodiment, by using the rotation matrix corresponding to the roll angle and pitch angle of the sensor to transform the coordinates of the accumulated point cloud, and filtering out points whose z-axis coordinates are less than the specified value, the influence of ground points on target detection can be reduced and the accuracy of target detection can be improved.
[0109] In an exemplary embodiment, performing a projection operation on a target frame set to obtain a target projection image includes:
[0110] S71, projecting the target frame set into a designated plane to obtain target projection information corresponding to the target frame set, wherein the target projection information is used to represent a pixel position to which each position point in the target frame set is projected and a pixel value corresponding to each position point, and the pixel value corresponding to each position point is a pixel value of each position point projected into a pixel point in the designated plane;
[0111] S72, generating a target projection image according to the pixel position to which each position point is projected and the pixel value corresponding to each position point.
[0112] When performing a projection operation on a target frame set, each position point in the target frame set can be projected into a specified plane respectively, thereby obtaining target projection information. Here, the target projection information is used to indicate the pixel position to which each position point in the target frame set is projected, and the pixel value corresponding to each position point. The pixel value corresponding to each position point is the pixel value of each position point projected into the specified plane. Here, the pixel value corresponding to each position point can be the label value of the aforementioned dirt label of each position point in the target frame set, or a value obtained by weighting the label value of the aforementioned dirt label of each position point in the target frame set, or a value determined in combination with other projection technologies, which is not limited in this embodiment.
[0113] A target projection image can be generated according to the pixel position to which each position point is projected and the pixel value corresponding to each position point. Here, the image area of the target projection image can be pre-set or determined according to the pixel positions to which all position points in the target frame set are projected. The width of the final projection image can be represented by width, and the height can be represented by height.
[0114] For example, the filtered LiDAR point cloud can be projected using Polar Stereographic Projection or other projection methods to project the spherical coordinates to the plane coordinates. The projected is the LiDAR spherical coordinate dirty mark. The formula for projecting the spherical coordinates to the plane coordinates can be as shown in Formula (4) and Formula (5):
[0115] θ=π-θ (4)
[0116]
[0117] Among them, radius is the projection radius; width is the width of the final projected image; height is the height of the final projected image.
[0118] The projection position can be calculated using formula (6) and formula (7):
[0119]
[0120] Among them, R is the polar coordinate radius after projection, θ is the polar coordinate angle after projection, and x and y are the x-axis coordinate and y-axis coordinate after projection.
[0121] Through this embodiment, by projecting each position point in the accumulated point cloud onto a designated plane, and generating a projection image based on the projection information of each position point, the convenience of generating a projection image can be improved.
[0122] In an exemplary embodiment, generating a target projection image according to the pixel position to which each position point is projected includes:
[0123] S81, when there is a position point set in the target frame set that is projected to the same pixel position, determine the average value of the pixel values corresponding to the position points in the position point set as the pixel value of the pixel position projected to the position point set in the target projection image;
[0124] S82, determining pixel values corresponding to other position points in the target frame set except for the position points in the position point set as pixel values of pixel positions to which the other position points in the target projection image are projected;
[0125] S83, when there is a group of pixel positions whose pixel values are not determined in the image area of the target projection image, determine the designated pixel value as the pixel value of the group of pixel positions in the target projection image.
[0126] For the target frame set, it may contain some position points projected to the same pixel position, or some position points whose projected positions are different from other position points. In the projected image, there may be some pixel positions to which no position points are projected, and the pixel values of the pixel positions in different situations can be determined respectively.
[0127] If there is a position point set in the target frame set that is projected to the same pixel position, the maximum value, minimum value or median value of the pixel values corresponding to the position points in the position point set can be determined as the pixel value of the pixel position in the target projection image to which the position point set is projected. Optionally, in order to improve the ability of the pixel value to characterize the state of the corresponding position point, the average value of the pixel values corresponding to the position points in the position point set can be determined as the pixel value of the pixel position in the target projection image to which the position point set is projected.
[0128] The pixel values corresponding to the other position points in the target frame set except the position points in the position point set are determined as the pixel values of the pixel positions projected to the other position points in the target projection image, and if there is a group of pixel positions with undetermined pixel values in the image area of the target projection image, the specified pixel value can be determined as the pixel value of a group of pixel positions in the target projection image. In this way, the pixel values of all pixel positions in the target projection image can be determined, thereby generating the target projection image.
[0129] For example, when there are multiple point clouds projected to the same position, the mean of all positions can be calculated as the pixel value of the corresponding pixel position on the projected image, as shown in formula (8):
[0130]
[0131] Among them, mean (x,y) is the mean value of dirt at the projection image position (x, y); tag i is the degree of dirtiness of the image point, and size is the number of point clouds projected to the projection image position (x, y). In addition, the positions without point cloud projection can also be numerically filled.
[0132] Through this embodiment, by generating a target projection image according to the pixel position to which each position point is projected, the accuracy of acquiring the target projection image can be improved.
[0133] In an exemplary embodiment, the target frame set is obtained by the target sensor detecting the sensor housing surface in one of a plurality of continuous time intervals. The classification result corresponding to each of the plurality of continuous time intervals is a confidence level output by the target classification model corresponding to each of the plurality of preset categories.
[0134] Correspondingly, after inputting the target projection image into the trained target classification model to obtain the target classification result, the above method further includes:
[0135] S91, performing smoothing processing on multiple confidences corresponding to the same preset category in the classification results corresponding to each time interval to obtain a smoothed confidence corresponding to each preset category;
[0136] S92: Determine, among the plurality of preset categories, a preset category corresponding to the highest smoothing confidence as a target category corresponding to the plurality of continuous time intervals.
[0137] In this embodiment, in order to increase the reliability of the inference results (i.e., the prediction results of the classification model), the confidences of the classification results corresponding to multiple continuous time intervals can be smoothed: first, multiple confidences corresponding to the same preset category in the classification results corresponding to each time interval are smoothed to obtain the smoothed confidences corresponding to each preset category, and then, among the multiple preset categories, the preset category with the highest corresponding smoothed confidence is determined as the target category corresponding to the multiple continuous time intervals.
[0138] For example, when using the trained model for inference deployment, you can connect to the lidar, process the lidar data in real time in a similar way to the above, and use the data processed by the model to obtain the inference results of the model. Considering that rain is a continuous process, it is necessary to smooth the continuous inference results within a certain time interval to increase the reliability of the inference results. Conf can be used to represent the confidence of rain at t moments, where Conf = {conf1,conf2,…,conf t},conf t =β*conf t-1 +(1-β)θ t , β = (t-1) / t is the preset weighting coefficient, θ t is the confidence that it will rain at time t, conf t is the confidence after exponential sliding average at time t.
[0139] The confidence of rain in a specified time interval can be smoothed using exponential moving average, as shown in formula (9):
[0140]
[0141] Wherein, thres is the preset confidence threshold of rain.
[0142] Through this embodiment, by performing smoothing processing on multiple classification results to obtain smoothed classification results, the credibility of the model output result can be enhanced.
[0143] In an exemplary embodiment, the method further comprises:
[0144] S101, performing a training sample acquisition operation multiple times based on a reference sensor to obtain a training image sample set, wherein the reference sensor and the target sensor are sensors of the same model, and the labeling information of each training image sample in the training image sample set is used to indicate one of a plurality of preset categories;
[0145] S102, performing model training on the target classification model to be trained based on each training image sample and the annotation information of each training image sample to obtain a trained target classification model.
[0146] In this embodiment, a selected sensor model (i.e., reference sensor) can be used to obtain training samples for classification model training: the training sample collection operation is performed multiple times based on the reference sensor to obtain a training image sample set. Here, the selected model is the model of the target sensor. Using the same model of sensor to collect training samples can improve the target detection effect of the trained classification model in actual scenes. The annotation information of each training image sample is used to indicate a preset category of multiple preset categories. The annotation here can be manual, machine combined with manual, or other annotation methods. In this embodiment, the annotation method of the training image samples is not limited.
[0147] In this embodiment, the training sample acquisition operation includes: obtaining a reference frame set obtained by detecting the sensor shell surface of the reference sensor through the reference sensor, wherein each reference frame in the reference frame set is a single frame point cloud obtained by a detection performed by the reference sensor; performing a projection operation on the reference frame set to obtain a training image sample. The manner of obtaining the reference frame set and performing the projection operation is similar to that in the aforementioned embodiment, and will not be described in detail here.
[0148] Based on each training image sample and the annotation information of each training image sample, the target classification model to be trained can be trained to obtain a trained target classification model. The method of training the classification model can refer to related technologies and is not limited in this embodiment.
[0149] For example, see Figure 5 In order to use LiDAR and deep neural network for rain detection and detect rain scenes and non-rain scenes, the training data can be obtained in a similar way as above (i.e., data collection): connect the selected model of LiDAR, use the connected LiDAR to collect data, and obtain positive and negative samples, where the positive sample is rain data and the negative sample is non-rain data. Here, the data collection steps include point cloud accumulation, point cloud filtering, and stereographic plane projection.
[0150] In addition to data collection, you can also select a model (i.e., a rain classification model), a loss function, an optimizer, and a data enhancement method: select a model to train a rain classification model. Here, you can select a suitable backbone network, such as the ResNet series, EfficientNet series, and other models. The output categories are 2 categories, rain or no rain; use random rotation, random upside down flip, random left-right flip, and random erasure to enhance the data; select binary cross entropy as the loss function; select Adam as the optimizer, and use the above configuration to train the model. For the trained model, you can evaluate the model performance. If the model evaluation indicators are met, output the model. If the evaluation indicators are not met, you can perform data analysis to supplement the data and adjust the parameters to retrain the model.
[0151] Here, the above-mentioned lidar rain detection solution based on deep neural network can effectively perform rain detection, that is, effectively detect rain scenes such as light rain, moderate rain and heavy rain. In addition, it can also effectively detect non-rain scenes such as dust, grass clippings, mud, etc., as long as the data collection and output categories of the corresponding scenes are increased.
[0152] See also Figure 6 When deploying rain detection applications, you can first connect to the lidar to obtain radar dirt data, that is, point cloud data; accumulate, filter and stereographically project the acquired point cloud data, then perform model inference and output the results; to enhance the credibility of the model output results, synthesize multiple inference results and output the final results.
[0153] Through this embodiment, a sensor of a specified model is used to collect training image samples, and the obtained training image samples are used to train the classification model, which can improve the target detection effect of the trained classification model in actual scenes.
[0154] It should be noted that, for the aforementioned method embodiments, for the sake of simplicity, they are all expressed as a series of action combinations, but those skilled in the art should be aware that the present application is not limited by the described order of actions, because according to the present application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the present application.
[0155] According to another aspect of the embodiment of the present application, a target detection device for a lawn mower is also provided, and the device is used to implement the target detection method for a lawn mower provided in the above embodiment, which has been described and will not be repeated. As used below, the term "module" can implement a combination of software and / or hardware for a predetermined function. Although the device described in the following embodiments is preferably implemented in software, the implementation of hardware, or a combination of software and hardware, is also possible and conceivable.
[0156] Figure 7 is a structural block diagram of a target detection device for a lawn mower according to an embodiment of the present application, such as Figure 7 As shown, the device comprises:
[0157] An acquisition unit 702 is used to acquire a target frame set obtained by detecting a sensor housing surface of a target sensor through a target sensor, wherein each target frame in the target frame set is a single frame point cloud obtained by performing one detection by the target sensor, and the target sensor is a sensor configured on the lawn mower for generating a point cloud;
[0158] The first execution unit 704 is used to perform a projection operation on the target frame set to obtain a target projection image, wherein the projection operation is an operation of projecting onto a specified plane;
[0159] The input unit 706 is used to input the target projection image into the trained target classification model to obtain a target classification result, wherein the target classification model is a classification model used to distinguish multiple preset categories, and the target classification result is used to indicate a preset category among multiple preset categories that matches the target frame set.
[0160] Through the embodiments of the present application, a target frame set is obtained by detecting the surface of a sensor shell of a target sensor through a target sensor, wherein each target frame in the target frame set is a single frame point cloud obtained by a detection by the target sensor, and the target sensor is a sensor configured on the lawn mower for generating a point cloud; a projection operation is performed on the target frame set to obtain a target projection image, wherein the projection operation is an operation of projecting onto a specified plane; the target projection image is input into a trained target classification model to obtain a target classification result, wherein the target classification model is a classification model for distinguishing multiple preset categories, and the target classification result is used to indicate a preset category among multiple preset categories that matches the target frame set, thereby solving the problem of large device size and high difficulty in device design in the target detection method in the related art.
[0161] In an exemplary embodiment, the acquiring unit includes:
[0162] A first acquisition module is used to acquire multiple groups of echo signals obtained by transmitting multiple detection signals to the surface of the sensor shell through the target sensor, wherein each group of echo signals in the multiple groups of echo signals includes two echo signals corresponding to the detection signal transmitted once;
[0163] The fusion module is used to fuse the two echo signals in each group of echo signals into a corresponding single-frame point cloud to obtain a target frame set.
[0164] In an exemplary embodiment, one echo signal in each group of echo signals is an echo signal reflected by the surface of the sensor housing, and the other echo signal is an echo signal reflected by the detected target.
[0165] In an exemplary embodiment, the fusion module includes:
[0166] The first determination subunit is used to determine the single-frame point cloud corresponding to each group of echo signals based on the signal difference at the same position point of two echo signals in each group of echo signals, and obtain a target frame set, wherein in the target frame set, the value of a position point corresponds to the signal difference between the two echo signal position points corresponding to the position point.
[0167] In an exemplary embodiment, each position point in each target frame is represented by a set of position parameters, and the value of each position point in each target frame is used to represent the degree of occlusion of the corresponding position point;
[0168] The above device also includes:
[0169] The second execution unit is configured to perform a filtering operation on the target frame set based on at least one position parameter to obtain a filtered target frame set, wherein the at least one position parameter belongs to a group of position parameters.
[0170] In an exemplary embodiment, the second execution unit includes at least one of the following:
[0171] The first filtering module is used to filter out the position points in the target frame set whose corresponding modulus length is not a specified value, and obtain the target frame set after modulus length filtering;
[0172] The second filtering module is used to filter out the position points in the target frame set whose corresponding polar angles are greater than or equal to the specified angle, so as to obtain the target frame set after polar angle filtering.
[0173] In an exemplary embodiment, a set of position parameters includes three axis coordinates in a rectangular coordinate system; the tilt angle of the target sensor includes a roll angle and a pitch angle;
[0174] The second execution unit includes:
[0175] A transformation module, used to transform the coordinate values of the three axis coordinates of each position point in the target frame set using a first rotation matrix and a second rotation matrix to obtain a target frame set after coordinate transformation, wherein the first rotation matrix is a rotation matrix determined according to a roll angle, and the second rotation matrix is a rotation matrix determined according to a pitch angle;
[0176] The third filtering module is used to filter out the position points whose corresponding z-axis coordinates are less than a preset value in the target frame set after coordinate transformation, so as to obtain the target frame set after coordinate filtering.
[0177] In an exemplary embodiment, the first execution unit includes:
[0178] A projection module is used to project the target frame set into a specified plane to obtain target projection information corresponding to the target frame set, wherein the target projection information is used to represent the pixel position to which each position point in the target frame set is projected and the pixel value corresponding to each position point, and the pixel value corresponding to each position point is the pixel value of each position point projected into the pixel point in the specified plane;
[0179] The generation module is used to generate a target projection image according to the pixel position to which each position point is projected and the pixel value corresponding to each position point.
[0180] In an exemplary embodiment, the generation module includes:
[0181] A second determining subunit is used to determine, when there is a position point set in the target frame set that is projected to the same pixel position, an average value of pixel values corresponding to the position points in the position point set as the pixel value of the pixel position projected to the position point set in the target projection image;
[0182] A third determining subunit is used to determine the pixel values corresponding to the other position points in the target frame set except the position points in the position point set as the pixel values of the pixel positions to which the other position points in the target projection image are projected;
[0183] The fourth determining subunit is configured to determine the designated pixel value as the pixel value of a group of pixel positions in the target projection image when there is a group of pixel positions whose pixel values are not determined in the image area of the target projection image.
[0184] In an exemplary embodiment, the target frame set is obtained by detecting the sensor shell surface by the target sensor in one of a plurality of continuous time intervals; the classification result corresponding to each of the plurality of continuous time intervals is a confidence level output by the target classification model and corresponding to each of a plurality of preset categories;
[0185] The above device also includes:
[0186] A processing unit, used for smoothing multiple confidences corresponding to the same preset category in the classification results corresponding to each time interval to obtain a smoothed confidence corresponding to each preset category;
[0187] The determination unit is used to determine the preset category with the highest corresponding smoothing confidence among the multiple preset categories as the target category corresponding to the multiple continuous time intervals.
[0188] In an exemplary embodiment, the above device further includes:
[0189] an acquisition unit, configured to perform a training sample acquisition operation multiple times based on a reference sensor to obtain a training image sample set, wherein the reference sensor and the target sensor are sensors of the same model, and the labeling information of each training image sample in the training image sample set is used to indicate one of the plurality of preset categories;
[0190] A training unit, used for performing model training on a target classification model to be trained based on each training image sample and the annotation information of each training image sample, to obtain a trained target classification model;
[0191] The training sample collection operation includes:
[0192] Acquire a reference frame set obtained by detecting a sensor shell surface of a reference sensor through a reference sensor, wherein each reference frame in the reference frame set is a single frame point cloud obtained by performing one detection by the reference sensor;
[0193] Perform a projection operation on the reference frame set to obtain a training image sample.
[0194] In an exemplary embodiment, the acquiring unit includes:
[0195] The second acquisition module is used to acquire a target frame set obtained by detecting the laser radar shell surface of the multi-dimensional laser radar through a multi-dimensional laser radar, wherein the target sensor is a multi-dimensional laser radar.
[0196] In an exemplary embodiment, the multi-dimensional laser radar is a three-dimensional laser radar.
[0197] In an exemplary embodiment, the target sensor is an integrated sensing device, and the target sensor is also used for at least one of the following: navigation, positioning, and obstacle avoidance.
[0198] In an exemplary embodiment, a plurality of preset categories are used to represent at least one of the following targets: rain, hail, snow, sandstorm, fog.
[0199] In an exemplary embodiment, the target sensor is used for rain detection, navigation, positioning, and obstacle avoidance.
[0200] In an exemplary embodiment, the sensor housing surface is planar and / or curved.
[0201] In an exemplary embodiment, the target sensor is disposed at an angle to the horizontal plane at a front region of the top of the lawn mower.
[0202] It should be noted that the above modules can be implemented by software or hardware. For the latter, it can be implemented in the following ways, but not limited to: the above modules are all located in the same processor; or the above modules are located in different processors in any combination.
[0203] According to another aspect of the embodiments of the present application, a computer-readable storage medium is provided, in which a computer program is stored, wherein the computer program is configured to execute the steps of any of the above method embodiments when run.
[0204] In an exemplary embodiment, the computer-readable storage medium may include, but is not limited to, various media that can store computer programs, such as a USB flash drive, a read-only memory (ROM), a random access memory (RAM), a mobile hard disk, a magnetic disk or an optical disk.
[0205] According to one aspect of the present application, a computer program product is provided, the computer program product comprising a computer program / instruction, the computer program / instruction comprising a program code for executing the method shown in the flowchart. Figure 8 The computer program can be downloaded and installed from the network through the communication part 809, and / or installed from the removable medium 811. When the computer program is executed by the central processor 801, various functions provided by the embodiments of the present application are executed. The serial numbers of the embodiments of the present application are only for description and do not represent the advantages and disadvantages of the embodiments.
[0206] refer to Figure 8 , Figure 8 It is a structural block diagram of a computer system of an optional electronic device according to an embodiment of the present application.
[0207] Figure 8 The computer system structure block diagram of the electronic device used to implement the embodiment of the present application is schematically shown. Figure 8As shown, the computer system 800 includes a central processing unit 801 (CPU), which can perform various appropriate actions and processes according to the program stored in the read-only memory 802 (ROM) or the program loaded from the storage part 808 to the random access memory 803 (RAM). Various programs and data required for system operation are also stored in the random access memory 803. The central processing unit 801, the read-only memory 802 and the random access memory 803 are connected to each other through a bus 804. The input / output interface 805 (I / O interface) is also connected to the bus 804.
[0208] The following components are connected to the input / output interface 805: an input section 806 including a keyboard, a mouse, etc.; an output section 807 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker; a storage section 808 including a hard disk, etc.; and a communication section 809 including a network interface card such as a local area network card, a modem, etc. The communication section 809 performs communication processing via a network such as the Internet. A drive 810 is also connected to the input / output interface 805 as needed. A removable medium 811, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 810 as needed so that a computer program read therefrom is installed into the storage section 808 as needed.
[0209] In particular, according to an embodiment of the present application, the process described in each method flow chart can be implemented as a computer software program. For example, an embodiment of the present application includes a computer program product, which includes a computer program carried on a computer readable medium, and the computer program contains a program code for executing the method shown in the flow chart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication part 809, and / or installed from the removable medium 811. When the computer program is executed by the central processor 801, various functions defined in the system of the present application are executed.
[0210] It should be noted that Figure 8 The computer system 800 of the electronic device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.
[0211] According to another aspect of the embodiments of the present application, there is also provided an electronic device, including a memory and a processor, wherein the memory stores a computer program, and the processor is configured to run the computer program to execute the steps in any one of the above method embodiments.
[0212] In an exemplary embodiment, the electronic device may further include a transmission device and an input / output device, wherein the transmission device is connected to the processor, and the input / output device is connected to the processor.
[0213] For specific examples in this embodiment, reference may be made to the examples described in the above embodiments and exemplary implementation modes, and this embodiment will not be described in detail herein.
[0214] According to another aspect of the embodiment of the present application, a lawn mower control method is also provided. The method embodiment provided in the embodiment of the present application can be used in a lawn mower (such as Figure 1 As shown), or a similar processing device is executed, which has been explained and will not be repeated here.
[0215] The lawn mower control method provided in this embodiment can be executed by the lawn mower alone, or by a user terminal (terminal device, which can be a mobile terminal, an integrated terminal, etc.) or a base station and a lawn mower together. Fig. 9 As shown, the process of the lawn mower control method may include the following steps:
[0216] S902, when there is water stain on the shell surface of the laser radar, obtaining a water signal obtained by the laser radar detecting the shell surface at least once, and determining the current weather based on the obtained water signal;
[0217] S904, controlling the lawn mower to perform corresponding actions according to the current weather; including: when it is determined that the current weather is bad weather, performing a shielding action on at least part of the lawn mower.
[0218] Similar to the aforementioned embodiment, the lawn mower control method in this embodiment can be applied to the scenario of target detection for a lawn mower or similar equipment. The lawn mower can be provided with at least one laser radar, and the environment in which the lawn mower is located can be detected by the provided laser radar to determine whether a water signal on the shell surface is detected. During the operation of the lawn mower (the lawn mower can be in a working state or an idle state, can be in a moving state or a stationary state, as long as the laser radar can be powered and the acquired laser radar data can be parsed), the laser radar can detect the shell surface and obtain the laser radar data; by parsing the laser radar data, the water signal on the above shell surface can be obtained.
[0219] It should be noted that the method of detecting the shell surface by laser radar is similar to the method of detecting the sensor shell surface by the target sensor in the aforementioned embodiment. In this case, the target sensor may include a laser radar, and correspondingly, the sensor shell surface of the target sensor may include the shell surface of the laser radar. Optionally, other methods may be used to detect the shell surface by laser radar to obtain a corresponding water signal. The water signal can be represented by a set of specified parameters, and the above set of specified parameters can be used to indicate whether there is water stain on the shell surface of the laser radar and the state of the water stain, which may include but is not limited to at least one of the following: the shape of the water, the area covered by the water, the position covered by the water, etc., and may also include other parameters that can characterize the water stain on the shell surface.
[0220] In this embodiment, the lawn mower can detect the environment in which the lawn mower is located by setting at least one laser radar to determine the current weather. Here, at least one laser radar can be set at any position of the lawn mower that can detect the environmental state, for example, the front, side, rear, etc. The positions of different laser radars can be different, and the detection areas of different laser radars can at least partially overlap or not completely overlap, which is not limited in this embodiment. The method of detecting bad weather by laser radar can be: parsing the point cloud data detected by the laser radar, and performing bad weather detection based on the parsed objects.
[0221] Optionally, the lawn mower may also be provided with a control module, for example, a first control module, and the lawn mower control method in this embodiment may be executed by the control module. The control module may obtain a water signal obtained by the laser radar detecting the shell surface. The above-mentioned acquisition operation may be implemented by: obtaining the laser radar data detected by the laser radar, and parsing the obtained laser radar data to obtain the corresponding water signal, thereby obtaining the water signal obtained by the laser radar detecting the shell surface.
[0222] Optionally, the control module can obtain laser radar data through the laser radar, and then obtain the state of the shell surface of the laser radar (that is, whether there are water stains on the shell surface of the laser radar). When water stains are detected on the shell surface, the current weather is determined based on the water signal obtained on the shell surface of the laser radar. The control module obtains the laser radar to detect the water signal on the shell surface periodically or by event triggering. If there are no water stains on the shell surface of the laser radar, the control module may not obtain the water signal. If there are water stains on the shell surface of the laser radar, the water signal obtained by the laser radar detecting its shell surface can be obtained; based on the obtained water signal, the current weather can be determined.
[0223] The determination of the current weather can be performed based on a water signal acquired once, or based on water signals acquired multiple times. Therefore, when there are water stains on the shell surface of the laser radar, the control module can obtain the water signal obtained by the laser radar detecting the shell surface at least once, and determine the current weather based on the acquired water signal. The current weather can be determined by determining what kind of weather it is (for example, rain, hail, snow, sunny, etc.), or by determining whether the current weather is a specified weather (for example, rain).
[0224] According to the current weather, the lawn mower can be controlled to perform corresponding actions. For example, if the current weather is sunny, the lawn mower can be controlled to continue mowing. For another example, if the current weather is raining, the lawn mower can be controlled to stop mowing. For another example, if the current weather is raining, the lawn mower can be controlled to continue mowing, return to the base station or a designated location according to the intensity of the rain, etc. The actions corresponding to different weather conditions can be preset or adjusted according to user needs.
[0225] In this embodiment, when it is determined that the current weather is bad weather, a shielding action is performed on at least part of the mower. The shielding action is performed on at least part of the mower to protect the laser radar or the entire mower from the influence of the bad external environment. Specifically, the above-mentioned action may include: controlling the mower to return to a base station that matches the mower, controlling the mower to walk to a specified position, and controlling a shielding member to shield at least part of the mower (laser radar, the body of the mower, etc.) so that at least part of the mower is shielded. The main working parts of the mower include a cutter disc, a laser radar, and a drive wheel. Continuous operation in a bad weather environment (such as rain) will cause the cutter disc to rust, the laser radar to be soaked in water, and the drive wheel to sink into mud, resulting in a reduction in the service life of the mower.
[0226] When the main body of the LiDAR is exposed to the external environment, it helps to detect weather conditions more accurately. However, when the main body of the LiDAR is exposed to the external environment, its rainproof performance will be affected, and the waterproof level of the LiDAR mounting components needs to be improved. The lawn mower happens to require a non-rainy working environment. Such conditions just meet the requirements for installing the LiDAR on the lawn mower to detect bad weather, without the need to add additional sealing structures to improve the waterproof level of the LiDAR, thereby reducing the manufacturing cost of the entire machine.
[0227] In this embodiment, it takes a certain amount of time to determine the current weather, and it also takes a certain amount of time for the weather to change. In order to ensure mowing efficiency, before determining the current weather, the lawn mower can be controlled to continue mowing according to a pre-set cutting path and cutting mode. Optionally, in order to improve the accuracy and timeliness of weather determination, the lawn mower can obtain weather information corresponding to the location of the lawn mower, and determine the current weather based on the obtained weather information, or determine the current weather based on the obtained weather information and the obtained water signal. There are many ways to obtain weather information corresponding to the location of the lawn mower, for example, directly obtaining weather information corresponding to the location of the lawn mower from the network (in this case, the lawn mower has an Internet connection function), and for example, obtaining weather information corresponding to the location of the lawn mower from a terminal device bound to the lawn mower. In this embodiment, there is no limitation on the way to obtain weather information corresponding to the location of the lawn mower.
[0228] Through the above steps, when there are water stains on the shell surface of the laser radar, a water signal obtained by the laser radar detecting the shell surface at least once is obtained, and the current weather is determined based on the obtained water signal; the lawn mower is controlled to perform corresponding actions according to the current weather; including: when it is determined that the current weather is bad weather, a shielding action is performed on at least part of the lawn mower, which can reduce the impact of weather factors on the operation of the lawn mower, ensure the safety of the lawn mower, and increase the service life of the lawn mower.
[0229] Before determining the current weather, the mower continues to mow according to the pre-set cutting path and cutting mode. There are two reasons for this setting: on the one hand, the laser radar has a short cycle for detecting rain, and in most cases the judgment result can be output within 30 seconds. In general, the judgment result will be output faster for heavy rain than for light rain. Therefore, before determining the current weather, the mower continues to mow according to the established mode and route, and the performance of the mower will not be affected by the rain. Instead, it can avoid the continuous start and stop of the walking motor and the cutting motor, ensuring the service life of the motor and not affecting the working efficiency of the mower. On the other hand, if the mower stops walking and cutting before determining the current weather, and the mower stays in place to continue to judge, if the mower stops at the location where it happens to be artificially precipitated (for example: a diversion pipe at a high place or a water container at a high place is dumped), this will cause the laser radar to continuously obtain water signals and misjudge the weather, and even cause the mower to fall into the mud, resulting in unfavorable results. In short, before determining the current weather, the lawn mower continues to mow according to a predetermined route and pattern, which can ensure the safety of the lawn mower, increase the service life of the lawn mower, and also ensure the mowing efficiency of the lawn mower.
[0230] In an exemplary embodiment, when the current weather is judged to be severe weather, a shielding action is performed on at least a portion of the lawn mower, including: when the current weather is judged to be severe weather, controlling the lawn mower to return to a base station matched with the lawn mower.
[0231] In this embodiment, if the current weather is judged to be bad weather, in order to avoid the bad weather causing damage to the lawn mower, the lawn mower can be controlled to return to the base station that matches the lawn mower. Usually, the base station is set in a sheltered space such as indoors to avoid short circuits, corrosion, dust accumulation, etc. caused by environmental factors such as sunlight, rain, wind and snow, thereby shortening the service life of the lawn mower. In addition, the matching relationship between the base station and the lawn mower can be pre-set, and there is no need to select a suitable parking position through some algorithms, thereby improving the timeliness of lawn mower control.
[0232] Here, bad weather refers to weather that may affect the operation of the lawn mower, such as heavy snow, freezing, low temperature, strong wind, high temperature, heavy rain and continuous rain, etc. Bad weather can be a custom configuration item (that is, bad weather can be weather specified by configuration information), which can correspond to the detected specified object or the object parameter of the detected specified object (for example, quantity, area, frequency, etc.), that is, when the specified object is detected or the object parameter of the detected specified object meets the set detection conditions, it is determined that the current weather is bad weather. It can also be other weather determination methods. Exemplarily, bad weather includes at least one of rain, hail, snow and fog. Here, by limiting the bad weather to be detected, the bad weather that can affect the operation of the lawn mower can be detected, thereby improving the rationality of the operation control of the lawn mower.
[0233] Optionally, the base station that controls the lawn mower to return can be pre-set, or it can be a corresponding base station selected from the area map (which can be the area address set on the lawn mower) according to the set selection rules after determining that the current weather is bad weather. The lawn mower can be charged and set up through this base station. At the same time, the base station can also provide a shielding function for the lawn mower.
[0234] Through this embodiment, when it is determined that the current weather is inclement weather, the lawn mower is controlled to return to a base station that matches the lawn mower, thereby improving the safety of the lawn mower's operation. At the same time, after the lawn mower returns to the base station, the lawn mower can be charged and maintained, so that the lawn mower can be better put into the next stage of work.
[0235] In an exemplary embodiment, when the current weather is judged to be bad weather, shielding at least part of the lawn mower includes: when the current weather is judged to be bad weather, controlling the lawn mower to move to a designated position so that at least part of the lawn mower is shielded. The designated position referred to in this embodiment can be under the eaves of a building, under an awning, or any sheltered place. After the lawn mower moves to the designated position, it can continue to perform the mowing action at the designated position, or it can stop the mowing action and stop somewhere.
[0236] In this embodiment, if the current weather is judged to be bad weather, in order to avoid the bad weather causing damage to the lawn mower, the lawn mower can be controlled to move to a designated location. Usually, the designated location will be set to a location that will not be affected by bad weather, such as a sheltered indoor location, or a location in the area without rain or hail, etc. The above-mentioned designated location can be pre-designated, such as a designated location in a canopy, or it can be selected from an area map (which can be an area address set on the lawn mower) according to a set selection rule after the current weather is judged to be bad weather. The designated location can be selected from a specified set of candidate locations, or it can be calculated in real time according to a selection algorithm, which is not limited in this embodiment.
[0237] Through this embodiment, when it is determined that the current weather is inclement weather, the lawn mower is controlled to move to a specified position, so that the lawn mower can be moved to a sheltered position along the shortest walking path, thereby improving the safety of the lawn mower operation and increasing the service life of the lawn mower.
[0238] In an exemplary embodiment, when the current weather is judged to be severe weather, a shielding action is performed on at least part of the lawn mower, including: when the current weather is judged to be severe weather, a shielding member is controlled to shield the laser radar, and the shielding member may be arranged on the lawn mower or outside the lawn mower. When the shielding member is arranged on the lawn mower, a control unit may be used to control the shielding member to shield the laser radar under severe weather conditions, and to be retracted under non-severe weather conditions to expose the laser radar so that the lawn mower can work normally; when the shielding member is arranged outside the lawn mower, a control unit may be used to control the shielding member to move to the position of the laser radar under severe weather conditions to shield the laser radar, and to leave the laser radar under non-severe weather conditions to remove the shielding of the laser radar.
[0239] In an exemplary embodiment, when the current weather is judged to be bad weather, a shielding action is performed on at least part of the lawn mower, including: when the current weather is judged to be bad weather, a shielding member is controlled to shield the body of the laser radar, and the shielding member can be set on the lawn mower or outside the lawn mower. When the shielding member is set on the lawn mower, a control unit can be used to unfold the shielding member to form a shield for the laser radar or retract the shielding member to remove the shielding for the laser radar; when the shielding member is set outside the lawn mower, the shielding member can be a movable shielding shed. When it is judged that the current weather is bad weather, the shielding shed can be moved to the outside of the lawn mower to form a shield for the lawn mower. When it is judged that the current weather is not bad weather, the shielding shed can leave the lawn mower so that the lawn mower can work normally.
[0240] In an exemplary embodiment, the method further comprises: in the case where it is determined that the current weather is bad weather, the lawn mower can be controlled to stop mowing, wherein the cutting assembly of the lawn mower is lifted after the mowing is stopped.
[0241] The lawn mower also includes a cutting assembly for performing mowing work, which can be located at the front side of the lawn mower. The cutting assembly can include a blade or other cutting assembly capable of cutting. When the lawn mower is working, the cutting assembly can extend outward for a distance, and after the work is completed, the cutting assembly can be retracted for a distance. The cutting component can move vertically up and down to improve the cutting of targets at different heights, and at the same time, it can avoid being blocked or unnecessary collisions when the target is not cut due to being too low. Exemplarily, the cutting assembly can be a knife disc.
[0242] If the current weather is judged to be bad weather, if the cutting component continues to perform the mowing work, it may cut grass covered with rain, fog, snow, etc., which will not only damage the cutting component, but also increase the weight of the collected grass if the lawn mower has a collection function, and may also make cleaning difficult. In this regard, the lawn mower can be controlled to stop mowing when the current weather is judged to be bad weather. Here, controlling the lawn mower to stop mowing can include controlling the cutting component to stop working. The way to control the cutting component to stop working can be achieved by controlling the cutting component to stop rotating, stopping power supply to the mowing component, etc.
[0243] Optionally, it takes a while for the cutting assembly to stop working. For example, it takes a while for the cutting assembly to stop from high-speed rotation. In order to avoid damage to the cutting assembly due to sudden stop (for example, circuit burnout), although a certain resistance can be added to increase the speed of stopping the cutting assembly, the resistance should not be too large. In order to ensure the safe operation of the cutting assembly, the cutting assembly can be controlled to be lifted when the current weather is judged to be bad weather. Correspondingly, the cutting assembly of the lawn mower is lifted after the mowing work is stopped. Lifting the cutting assembly can not only improve the convenience of moving the lawn mower, but also prevent the cutting assembly from being damaged due to insufficient cutting force when cutting grass stained with rain, fog, snow, etc. before the cutting assembly stops working.
[0244] Through this embodiment, when it is determined that the current weather is inclement weather, the lawn mower can be controlled to stop mowing. After stopping mowing, the cutting component of the lawn mower is lifted up, which can reduce the risk of damage to the cutting component, improve the convenience of moving the lawn mower, and thus increase the service life of the cutting component.
[0245] In an exemplary embodiment, controlling the lawn mower to perform corresponding actions according to the current weather includes: when it is determined that the current weather is not severe weather, controlling the lawn mower to continue mowing according to a preset cutting path and cutting mode.
[0246] Similar to the aforementioned embodiment, the lawn mower control method in this embodiment can be applied to the scenario of target detection for a lawn mower or similar equipment. The lawn mower may include a laser radar, the number and setting position of the laser radar, the characterization parameters of the water signal, and the way in which the lawn mower obtains the laser radar to detect the shell surface to obtain the water signal is similar to the aforementioned embodiment and will not be repeated here. During the operation of the lawn mower, the laser radar can detect the shell surface and obtain the laser radar data; by parsing the laser radar data, the water signal on the above shell surface can be obtained.
[0247] Similar to the above-mentioned embodiment, when there are water stains on the shell surface of the laser radar, the water signal obtained by the laser radar detecting the shell surface at least once can be obtained, and the current weather can be determined based on the obtained water signal. Here, taking into account the different determination methods adopted, the current weather can be determined based on the water signal obtained by the laser radar detecting the shell surface once, or it can be determined based on the water signal obtained by the laser radar detecting the shell surface once; according to the current weather, the lawn mower can be controlled to perform corresponding actions, such as returning to a specified position, returning to a base station matching the lawn mower, continuing mowing, etc., which have been explained and will not be repeated here.
[0248] In this embodiment, controlling the lawn mower to perform corresponding actions according to the current weather may include: when it is determined that the current weather is not severe weather, controlling the lawn mower to continue mowing according to a pre-set cutting path and cutting mode. The manner of controlling the lawn mower to continue mowing is similar to that in the above-mentioned embodiment, which has been described and will not be repeated here. Exemplarily, severe weather includes at least one of rain, hail, snow, and fog.
[0249] Through the above steps, when there are water stains on the shell surface of the laser radar, a water signal obtained by the laser radar detecting the shell surface at least once is obtained, the current weather is judged based on the obtained water signal, and the lawn mower is controlled to perform corresponding actions according to the current weather; the lawn mower is controlled to perform corresponding actions according to the current weather, including: when it is determined that the current weather is not bad weather, the lawn mower is controlled to continue mowing according to a pre-set cutting path and cutting mode, which can avoid affecting the working progress of the lawn mower and ensure the mowing efficiency of the lawn mower.
[0250] Specifically, the use of laser radar for weather detection in this embodiment has a lower probability of misjudgment than the use of rain sensors for weather detection. For example, when a user pours a cup of water on the rain sensor, the resistivity of the rain sensor changes, generating an electrical signal, and non-bad weather is judged as bad weather based on the generation of the electrical signal, thereby controlling the lawn mower to return to the base station or to the designated location; the use of laser radar can avoid the above-mentioned misjudgment. Similarly, the user pours a cup of water on the surface of the laser radar. Since the distribution characteristics of the water signal formed by pouring a cup of water on the surface of the laser radar are different from the distribution characteristics of the water signal formed by rain on the surface of the laser radar, the laser radar can distinguish between the two, and therefore will not judge such a situation as bad weather. At this time, the lawn mower will not return to the base station or the designated location, but will continue mowing according to the pre-set cutting path and cutting mode.
[0251] In an exemplary embodiment, when there are water stains on the shell surface of a laser radar, a water signal obtained by the laser radar detecting the shell surface at least once is obtained, and the current weather is determined based on the obtained water signal, including: obtaining a target frame set of water signals obtained by the laser radar detecting the shell surface at least once, comparing the processed target frame set with a preset classification model, and determining the current weather according to the target classification result, wherein the preset classification model is classified according to different types of weather.
[0252] In this embodiment, in order to determine the current weather, the lawn mower can collect and store enough signals from the surface of the laser radar shell in different scenarios (for example, under trees, under eaves, etc.) and in bad weather in different regions (for example, rainy weather), and obtain a classification model after training. The trained classification model can identify bad weather and non-bad weather based on the target frame set of water signals detected. The training process of the classification model can be similar to that in the aforementioned embodiment, and it has been explained and will not be repeated here. Correspondingly, the above-mentioned classification model can be pre-configured on the lawn mower (which can also be a server or other device), that is, a preset classification model, which can be used to determine the current weather.
[0253] A target frame set of water signals obtained by at least one detection of the shell surface by the laser radar can be obtained. Here, each target frame in the target frame set is a single-frame point cloud obtained by one detection by the laser radar. The target frame set can include a set of point clouds obtained by multiple detections of the shell surface by the laser radar within a certain period of time. The detection cycle of the laser radar for detecting the shell surface can be pre-configured, and the multiple detections can be continuous multiple detections or non-continuous multiple detections. The process of obtaining the target frame set is similar to that in the aforementioned embodiment and will not be repeated here.
[0254] Optionally, in order to facilitate the use of the preset classification model, the obtained target frame set can be preprocessed, for example, a projection operation is performed on the target frame set to obtain a target projection image. Here, the process of performing the projection operation on the target frame set is similar to that in the aforementioned embodiment and is not described in detail. For the obtained target frame set, the target frame set can be processed and compared with the preset classification model, and the current weather can be determined according to the target classification result.
[0255] The weather category of the preset classification model can be a specific weather category, such as sunny, rainy, snowy, etc., or it can be bad weather and non-bad weather. The method of comparing the target frame set with the preset classification model after processing can be: input the target projection image into the preset classification model for comparison, and the comparison can be the target projection image and the preset projection images corresponding to different weather categories or other features to determine the weather category corresponding to the target frame set, that is, the weather category indicated by the target classification result. The process of inputting the target projection image into the preset classification model to obtain the classification result is similar to that in the aforementioned embodiment, and will not be repeated here.
[0256] Through this embodiment, by acquiring a target frame set of water signals obtained by at least one detection of the shell surface by a laser radar, the target frame set is processed and compared with a preset classification model, and the current weather is judged according to the target classification result, the accuracy and reliability of the current weather judgment can be improved.
[0257] In an exemplary embodiment, a target frame set of water signals obtained by at least one detection of the shell surface by a laser radar is obtained, the target frame set is processed and compared with a preset classification model, and the current weather is determined according to the target classification result, including: performing J detections on the shell surface of the laser radar within a detection cycle, if the judgment results of the first K detections in the J detections are that the current weather is bad weather, then the final weather is output as bad weather; if the judgment results of the first L detections in the J detections are that the current weather is bad weather, and the judgment result of the L+1th detection is that the current weather is not bad weather, then the next detection cycle is entered until the current weather is determined; if at least one detection result in the judgment results of the first L detections in the J detections is that the current weather is not bad weather, then the final current weather is not bad weather is output, wherein L<K≤J and J≥2.
[0258] In order to reduce the error in judging the current weather, the shell surface of the laser radar can be detected J times (J ≥ 2) in one detection cycle, and the current weather can be judged based on the detection results of different times. Optionally, in this embodiment, the method of judging the current weather based on the judgment results of J times of detection can be:
[0259] If the judgment results of the first K tests in the J tests are all that the current weather is bad weather, then the final weather is output as bad weather. For example, the shell surface is detected three times in one cycle, and the detection result is marked as 1 for rain, and marked as 0 for non-rain. When the three output results are 111, it is judged to be rainy weather.
[0260] If the judgment results of the first L detections in the J detections are all that the current weather is bad weather, and the judgment result of the L+1th detection is that the current weather is not bad weather, then the next detection cycle is entered until the current weather is determined. For example, when the results of the three outputs are 110, the next cycle is entered;
[0261] If at least one of the judgment results of the first L detections in the J detections shows that the current weather is not bad weather, then the final output weather is not bad weather. For example, when the results of three outputs are 100 or 000, it is determined that it is not raining.
[0262] Here, L<K≤J. The above judgment process can be understood as follows: if the judgment results of multiple consecutive tests are that the current weather is rainy, then it is judged to be rainy; if the judgment results of the first few tests are that the current weather is rainy, and the judgment result of the last test is that the current weather is not rainy, then the current weather cannot be accurately judged (the judgment result of the last test is non-rainy weather), and it is necessary to enter the next cycle and continue to judge; and if at least one of the judgment results of the first L tests is that the current weather is not bad weather, and the judgment result of the L+1th test is that the current weather is not bad weather (that is, the judgment results of multiple tests are that the current weather is not rainy weather, and the judgment result of the last test is that the current weather is not bad weather), then it can be judged that the current weather is not bad weather.
[0263] Through this embodiment, the shell surface of the laser radar is detected multiple times within one detection cycle, and the current weather is comprehensively determined based on the determination results of the multiple detections, which can improve the accuracy and reliability of the current weather determination.
[0264] In an exemplary embodiment, the number of detections within a detection cycle can be pre-set or adjusted according to user needs. In order to improve the timeliness of the current weather judgment and at the same time improve the accuracy of the current weather judgment, the set number of detections should not be too many or too few. The appropriate number of detections can be predicted based on simulation data or actual experimental data.
[0265] Optionally, in this embodiment, J=3, K=3, L=2 can be set, that is, the shell surface of the laser radar is detected three times within one detection cycle. If the judgment results of the three detections are all that the current weather is bad weather, then the final output that the current weather is bad weather is bad weather; if the judgment results of the first two detections among the three detections are all that the current weather is bad weather, and the judgment result of the third detection is that the current weather is not bad weather, then the next detection cycle is entered until the current weather is determined; if the judgment results of the first two detections among the three detections show that the current weather is not bad weather, and the judgment result of the last detection is that the current weather is not bad weather, then the final output that the current weather is not bad weather is output.
[0266] Through this embodiment, by setting an appropriate number of detection times within a cycle, the timeliness and accuracy of current weather determination can be improved.
[0267] In an exemplary embodiment, the duration of a detection cycle is negatively correlated with the temperature of the shell surface. Here, considering that the surface temperature of the laser radar shell surface has an effect on the evaporation rate of water (i.e., in a high temperature environment, the evaporation rate of water is accelerated; in a low temperature environment, the evaporation rate of water is slowed down), the duration of a detection cycle can be set based on the temperature of the shell surface, wherein the duration of a detection cycle is negatively correlated with the temperature of the shell surface.
[0268] Through this embodiment, the duration of a detection cycle is set according to the temperature of the shell surface, so that the efficiency of the current weather determination can be improved while ensuring the accuracy of the current weather determination.
[0269] In an exemplary embodiment, the corresponding relationship between the temperature of the shell surface and the duration of a set detection cycle can be preset, and the above corresponding relationship can be predicted based on simulation data or actual experimental data.
[0270] In this embodiment, the duration of the detection cycle can be set directly, or the duration of the detection cycle can be obtained based on the correspondence between the temperature of the shell surface and the duration of a detection cycle. Considering that the duration of the detection cycle should not be too long (too long will lead to untimely determination of the current weather) or too short (too short will lead to inaccurate determination of the current weather), a value range of the duration of a detection cycle can be set. For example, the duration of a detection cycle is 10-30S, that is, the duration of the detection cycle is directly set within 10-30S, and the duration of the detection cycle allowed by the shell surface temperature is within 10-30S.
[0271] Through this embodiment, by setting a value range for the duration of a detection cycle, the efficiency of the current weather determination can be improved while ensuring the accuracy of the current weather determination.
[0272] According to another aspect of the embodiment of the present application, a lawn mower control method is also provided. The method embodiment provided in the embodiment of the present application can be used in a lawn mower (such as Figure 1 As shown), or a similar processing device is executed, which has been explained and will not be repeated here.
[0273] The lawn mower control method provided in this embodiment can be executed by the lawn mower alone, or can be executed by the user end or the base station and the lawn mower together. Fig.10 As shown, the process of the lawn mower control method includes the following steps:
[0274] S1002, when there is water stain on the shell surface of the laser radar, obtaining water signals obtained by detecting the shell surface by the laser radar at least twice within a set total time period, and determining the current weather based on the obtained water signals;
[0275] S1004, controlling the lawn mower to perform corresponding actions according to the current weather; including: when it is determined that the current weather is inclement weather, performing a shielding action on at least part of the lawn mower.
[0276] Similar to the aforementioned embodiment, the lawn mower control method in this embodiment can be applied to the scenario of target detection for a lawn mower or similar equipment. The lawn mower may include a laser radar. The number and setting position of the laser radar, the characterization parameters of the water signal, and the way in which the lawn mower obtains the laser radar to detect the shell surface to obtain the water signal are similar to those in the aforementioned embodiment and will not be repeated here. During the operation of the lawn mower (the lawn mower can be in a working state or an idle state, can be in a moving state or a stationary state, as long as it can power the laser radar and parse the acquired laser radar data), the laser radar can detect the shell surface and obtain the laser radar data; by parsing the laser radar data, the water signal on the above-mentioned shell surface can be obtained.
[0277] Optionally, in order to ensure the accuracy of the current weather detection, the current weather can be determined based on the changes in the multiple water signals obtained. In this embodiment, when there are water stains on the shell surface of the laser radar, the water signal on the shell surface of the laser radar can be detected at least twice within the set total time period (that is, the water signal obtained by the laser radar detecting its shell surface at least twice); based on the changes in the obtained water signal, the current weather is determined, and the timing for determining the current weather can be after the current weather is determined based on a certain water signal obtained, or it can be at the end of the total time period.
[0278] In the process of acquiring water signals obtained by detecting the shell surface by the laser radar at least twice within a set total time period, the time interval between two adjacent acquisitions of water signals can be fixed (for example, acquiring a water signal every 5 seconds), or can be flexibly set based on the detected environmental parameters. When acquiring water signals at different times, the position of the lawn mower can be the same, that is, acquiring water signals obtained by detecting the shell surface by the laser radar at least twice in a stationary state, or it can be different, for example, adjusting the posture (position and / or attitude) of the lawn mower each time a water signal is acquired.
[0279] Within the set total time period, water signals obtained by at least two laser radar detections of the shell surface can be obtained. The set total time period can be a time period of a specified duration set based on the first detection of the water signal, for example, a time period of 20 seconds starting from the moment the water signal was first detected. Based on each water signal obtained, the current weather can be determined according to different conditions until the determination condition is met or the total time period ends. Here, the determination condition can be a condition for determining whether the current weather is bad weather or a certain type of weather, or it can be other conditions that are pre-set or adjusted according to user needs.
[0280] According to the current weather, the lawn mower can be controlled to perform corresponding actions. The actions corresponding to different weather conditions can be preset or adjusted according to user needs. The manner of controlling the lawn mower to perform corresponding actions according to the current weather is similar to that in the above-mentioned embodiment, which has been described and will not be repeated here.
[0281] Through the above steps, when there are water stains on the shell surface of the laser radar, water signals obtained by detecting the shell surface by the laser radar at least twice within the set total time period are obtained, and the current weather is determined based on the water signal obtained each time; the lawn mower is controlled to perform corresponding actions according to the current weather; including, when it is determined that the current weather is bad weather, shielding at least part of the lawn mower, which can improve the accuracy of the current weather judgment, avoid weather damage to the lawn mower or affect the normal operation of the lawn mower, ensure the safety of the lawn mower, and increase the service life of the lawn mower.
[0282] On a lawn mower, using LiDAR to detect rain has a greater advantage than using traditional rain sensors to detect rain. Rain sensors use the change in resistivity to detect the presence or flow of rain. When rain falls on the sensor surface, it changes its resistivity, thereby generating an electrical signal. By detecting the change in this electrical signal, the presence of rain is determined. The disadvantage of using a rain sensor to detect rain is that as long as there are water stains on the surface of the rain sensor, it will be judged as rain. It cannot distinguish between artificial water splashing or accidental water leakage, so it will be misjudged as rain. The lawn mower is not suitable for working in a rainy environment, which causes the lawn mower to stop mowing or return to the base station, resulting in reduced mowing efficiency. LiDAR can solve the above defects because it can judge whether the water signal on the surface of the LiDAR is naturally generated or caused by humans based on the water signal frame set and water signal changes on its surface. The detection accuracy is greatly improved, ensuring the mowing efficiency of the lawn mower. In addition, this embodiment obtains water signals obtained by detecting the shell surface by laser radar at least twice within the set total time period to judge the current weather. Obtaining two water signals can compare and analyze the difference between the two water signals, so as to accurately distinguish whether the water signal is caused by rain or human factors, reducing the probability of misjudgment.
[0283] In this embodiment, when it is judged that the current weather is rainy, a shielding action is performed on at least part of the lawn mower. The shielding action is performed on at least part of the lawn mower to protect the laser radar or the lawn mower as a whole from the influence of the harsh external environment. Specifically, the action of avoiding bad weather may include but is not limited to one of the following: controlling the lawn mower to return to a base station that matches the lawn mower, controlling the lawn mower to walk to a specified position, and controlling a shielding member to shield at least part of the lawn mower (the laser radar, the body of the lawn mower, etc.) so that at least part of the lawn mower is shielded. The main working parts of the lawn mower include a cutter disc, a laser radar, and a drive wheel. Continuous operation in a harsh weather environment (such as rain) will cause the cutter disc to rust, the laser radar to be soaked in water, and the drive wheel to sink into the mud, resulting in a reduction in the service life of the lawn mower.
[0284] When the main body of the LiDAR is exposed to the external environment, it helps to detect weather conditions more accurately. However, when the main body of the LiDAR is exposed to the external environment, its rainproof performance will be affected. The lawn mower just needs a non-rainy environment. Such conditions just meet the requirements for installing the LiDAR on the lawn mower to detect bad weather.
[0285] In an exemplary embodiment, before acquiring water signals obtained by detecting the shell surface by laser radar at least twice within a set total time period, the method further includes: setting the length of the total time period according to the temperature of the shell surface, wherein the length of the total time period is negatively correlated with the temperature of the shell surface.
[0286] Taking into account that the surface temperature of the lidar surface affects the evaporation rate of water (i.e., in a high temperature environment, the evaporation rate of water is accelerated; in a low temperature environment, the evaporation rate of water is slowed down), the length of the total time period can be set based on the temperature of the shell surface, where the length of the total time period is negatively correlated with the temperature of the shell surface.
[0287] Here, if the temperature of the shell surface is low and the water evaporates slowly, if the total time period is set to be short, for scenes without rain, the difference between the obtained water signals is small, and it is easy to be mistaken for rainy weather; if the temperature of the shell surface is high and the water evaporates quickly, even if the total time period is set to be short, the difference between the obtained water signals can still accurately characterize the current weather.
[0288] The correspondence between the shell surface temperature and the duration of the total time period can be preset, which can be represented by a correspondence table, a correspondence curve or other methods. According to the correspondence between the shell surface temperature and the duration of the total time period, the duration of the total time period is set to the duration corresponding to the current shell surface temperature.
[0289] Through this embodiment, the duration of the total time period is set according to the temperature of the shell surface, so that the efficiency of the current weather determination can be improved while ensuring the accuracy of the current weather determination.
[0290] In an exemplary embodiment, the correspondence between the temperature of the shell surface and the duration of the set total time period can be preset, and the above correspondence can be predicted based on simulation data or actual experimental data.
[0291] In this embodiment, taking into account the case of outdoor operation, the shell surface temperature is usually higher. When the shell surface temperature is 35-60°C, in order to avoid the influence of high temperature on detection accuracy, the set total time can be 10-30S (seconds). For example, when the shell surface temperature is 35°C, the set total time can be 30S. When the shell surface temperature is 60°C, the set total time can be 10S.
[0292] The detection time of water stains by laser radar is negatively correlated with the surface temperature of the shell of the laser radar. Under the condition of the same detection accuracy, the higher the temperature of the shell surface, the shorter the detection time required, and the lower the temperature of the shell surface, the longer the detection time required; while the shell surface temperature is positively correlated with the working time of the lawn mower. Generally speaking, after the laser radar has worked for 10 minutes, the shell surface temperature can reach 35°C in most cases (because the environment in which grass grows, the ambient temperature itself will not be too low). After testing, when the shell surface temperature reaches 35°C and the detection time reaches 10S, the accuracy of detecting bad weather is greater than 86%, and the longer the working time of the laser radar, the higher the shell surface temperature. When the working time reaches 30min, the shell surface temperature of the laser radar can reach 40°C in most cases. At this time, when the detection time is set to 10S, the accuracy of detecting bad weather is greater than 93%. When the shell surface temperature of the laser radar reaches 60°C and the detection time is set to 10S, the accuracy of detecting bad weather is greater than 99%. Therefore, for users, under long-term use conditions, when the detection time is set to 10S, the overall probability of misjudgment of bad weather detection is less than 10%, and the misjudgment is that non-bad weather is judged as bad weather. This misjudgment will not lose the service life of the mower or lidar, but will only make the mower perform some unnecessary actions to avoid bad weather, affecting the mowing efficiency of the mower. These situations are acceptable to users. If the detection time is less than 10S, the misjudgment rate of detection will rise sharply, and the loss of mower working efficiency will be great, affecting the user experience. After testing, when the temperature of the shell surface reaches 35℃ in most cases and the detection time reaches 30S, the accuracy of bad weather is greater than 96%. When the temperature of the shell surface can reach 40℃ in most cases and the detection time is set to 30S, the accuracy of bad weather detection is greater than 99%. Therefore, when the detection time reaches 30S, the detection accuracy is high enough to fully meet the inspection requirements. Increasing the detection time to more than 30S has a very weak effect on the detection accuracy, but it will increase the data processing load of the processing unit and affect the service life of the processing unit.
[0293] Through this embodiment, when the temperature of the shell surface of the laser radar is 35-60°C, the duration of the total time period is set to 10-30S, which can improve the efficiency of the current weather judgment while ensuring the accuracy of the current weather judgment.
[0294] In an exemplary embodiment, the total time period is divided into multiple sub-time periods, and the durations of different time periods in the multiple sub-time periods may be the same or different. The division criteria of the multiple sub-time periods may be pre-configured, for example, the total time period corresponding to the shell surface temperature change and the division criteria of the total duration may be configured, or the user may directly set the division criteria on the user side, for example, detection once every 5S, or other division criteria, which are not limited in this embodiment.
[0295] Correspondingly, a water signal obtained by detecting the shell surface by the laser radar at least twice within the set total time period is obtained, including: within the set total time period, a water signal obtained by detecting the shell surface by the laser radar once at the end of each sub-time period in multiple sub-time periods.
[0296] The total time period can be divided into multiple sub-time periods, and correspondingly, the water signal obtained by the laser radar detecting the shell surface is continuously obtained within the set total time period, including: within the set total time period, the water signal obtained by the laser radar detecting the shell surface is obtained once at the end of each sub-time period in multiple sub-time periods.
[0297] In this embodiment, starting from the moment when the water signal is detected for the first time, the water signal is acquired once at a certain interval (which may be the end moment of a sub-time period of multiple sub-time periods). The above-mentioned total time period and multiple sub-time periods can be characterized by acquiring a water signal once at a certain interval. For example, the total time period can correspond to the number of times the water signal is acquired, and the multiple sub-time periods can correspond to the time interval between two adjacent acquisitions of the water signal. The time interval between two adjacent acquisitions of the water signal can be the same or different.
[0298] Optionally, the end time of each sub-time period can be determined by setting a timer according to the duration of each sub-time period, and determining that each sub-time period has reached its end position when the timer's timing time is reached. Accordingly, starting from the moment when the water signal is first acquired, the timing time of the timer is set, and the timing time can be a specified fixed time or a time set based on the temperature of the shell surface; when the timing time of the timer is reached (i.e., the timer times out), a water signal is acquired once, and the current weather is determined. If the current weather cannot be determined, the timing time of the timer is reset; and so on, until the number of water signals acquired reaches the specified number.
[0299] Through this embodiment, within the set total time period, a water signal is acquired once every certain period of time (corresponding to a sub-time period), which can improve the efficiency and controllability of water signal acquisition.
[0300] In an exemplary embodiment, the durations of different time periods in the multiple sub-time periods may be the same or different. Considering the number of timers required to be set and the complexity of the current weather determination, in this embodiment, the durations of the various sub-time periods may be set equal, for example, the duration of each sub-time period is 5S.
[0301] Through this embodiment, the duration of each sub-time end is set to be equal, which can reduce the number of timers that need to be set and reduce the complexity of current weather determination.
[0302] In an exemplary embodiment, in order to improve the timeliness of the current weather determination and the accuracy of the current weather determination, the total time period set should not be too long or too short, and the divided time periods should also be appropriate. In this embodiment, when the total time period is 15 seconds, the total time period is divided into three sub-time periods, each of which is 5 seconds long. The total time period of 15 seconds can be set according to the temperature of the shell surface, or it can be a preset fixed value, or it can be set in other ways.
[0303] It should be noted that the duration of different time periods in the multiple sub-time periods may be the same, the number of sub-time periods included in the multiple sub-time periods may be three, and the duration of a single sub-time period may be the duration of the total time period / 3. When the duration of the total time period is 15S, the duration of each sub-time period is 5S. When the duration of the total time period is other values, it may be divided into three sub-time periods in the same or similar division manner.
[0304] Through this embodiment, when the total time period is 15S, by dividing each sub-time period into 5S, the efficiency and controllability of water signal acquisition can be improved, thereby improving the timeliness and accuracy of current weather judgment.
[0305] In an exemplary embodiment, water signals obtained by detecting the shell surface by the laser radar at least twice within a set total time period include: obtaining water signals obtained by detecting the shell surface by the laser radar at different postures within the set total time period.
[0306] There may be many reasons for water stains on the shell surface of the LiDAR, such as rain, continuous water spraying (for example, watering the grass), etc. Therefore, in addition to rain, there are other situations in which water signals can be continuously detected from the shell surface of the LiDAR. In order to ensure the accuracy of the current weather judgment, the water signals obtained by the LiDAR detecting the shell surface at different positions can be combined to make the current weather judgment.
[0307] Correspondingly, within the set total time period, the water signal obtained by the laser radar detecting the shell surface at different postures (positions and / or attitudes) can be obtained. After obtaining the water signal at a certain posture, if it is necessary to continue to obtain the water signal, the posture of the laser radar (at least one of the position and attitude) can be adjusted, and the water signal can be re-acquired at the new posture. Adjusting the posture of the laser radar can be achieved by adjusting the posture of the lawn mower, or by adjusting the posture of the laser radar.
[0308] Through this embodiment, the water signal obtained by the laser radar detecting the shell surface at different postures is obtained, which can improve the accuracy of the current weather judgment.
[0309] In an exemplary embodiment, obtaining a water signal obtained by a laser radar detecting a shell surface in different postures includes: obtaining a water signal obtained by a laser radar detecting a shell surface when a lawn mower moves to different positions, wherein the posture of the laser radar is different when the lawn mower moves to different positions.
[0310] When the lawn mower continues to move, its position will change, and the position and posture of the laser radar will also change. In this embodiment, the water signal obtained by the laser radar detecting the shell surface when the lawn mower moves to different positions can be obtained. The movement of the lawn mower can be the movement of the lawn mower when it is mowing, or it can be the movement of the lawn mower to obtain the corresponding water signals in different postures.
[0311] The timing for acquiring the water signal can be that the time interval between two adjacent acquisitions of the water signal is a specified time interval, or, when two adjacent water signals are acquired, the distance between the positions of the lawn mower is a specified distance, or other acquisition timings can be used, as long as it can be ensured that the acquired water signal is the water signal obtained by the laser radar detecting the shell surface when the lawn mower moves to different positions.
[0312] Acquiring the water signal obtained by the laser radar detecting the shell surface at different postures within the set total time can eliminate some misjudgments, such as judging a non-raining situation as a raining situation. The following scenarios will cause the above-mentioned misjudgments: the lawn mower moves under a high drain pipe, the lawn mower moves under a leaking water container, the lawn mower moves under a swing with an overturned beverage bottle placed on the swing. These situations will cause the laser radar to continuously detect water signals on the shell surface. If the position is not moved, the laser radar will misjudge this situation as rain, thereby controlling the lawn mower to perform unnecessary actions, such as returning to the base station, thereby reducing mowing efficiency.
[0313] Through this embodiment, the water signal obtained by the laser radar detecting the shell surface when the lawn mower moves to different positions is obtained, which can improve the convenience of current weather judgment while ensuring the working efficiency of the lawn mower.
[0314] In an exemplary embodiment, a signal detection condition may be set, that is, a condition satisfied when the laser radar detects the shell surface. The signal detection condition may be a posture condition, which may include at least one of the following: a direction different from the direction when the water signal was acquired last time; a position different from the position when the water signal was acquired last time; and a distance between the position and the position when the water signal was acquired last time is a specified distance.
[0315] Correspondingly, in this embodiment, the water signal obtained by the laser radar detecting the shell surface in different postures is obtained, including: when the lawn mower is in a working state, the water signal obtained by the laser radar detecting the shell surface for the first time is obtained; the lawn mower is controlled to move or rotate to a detection posture that meets the signal detection conditions, and the water signal obtained by the laser radar detecting the shell surface when the lawn mower is in the detection posture that meets the signal detection conditions is obtained.
[0316] When the lawn mower is in operation, the water signal obtained by the laser radar detecting the shell surface is obtained for the first time, that is, the water signal obtained by the laser radar detecting the shell surface is obtained for the first time (the process of obtaining the water signal from nothing to something). At this time, the lawn mower can be controlled to move or rotate to a detection posture that meets the signal detection conditions, and when the lawn mower moves or rotates to a detection posture that meets the signal detection conditions, the water signal obtained by the laser radar detecting the shell surface is obtained. After each water signal is obtained, the current weather can be determined. If it is necessary to obtain a water signal, the posture of the lawn mower can be adjusted in the same or similar manner as described above.
[0317] Through this embodiment, the detection posture of the laser radar for detecting water signals is controlled by signal detection conditions, so that the convenience of controlling the lawn mower can be improved. At the same time, the accuracy of the current weather judgment can be improved by using the water signals detected under different detection postures to judge the current weather.
[0318] In an exemplary embodiment, a determination condition for determining whether to continue to acquire a water signal may be set, for example, a first determination condition, which may be a condition for determining that the water stain on the surface of the laser radar is not caused by bad weather. Correspondingly, based on each acquired water signal, the current weather is determined, including: determining the current weather according to the first determination condition until the first determination condition is satisfied or the total time period ends.
[0319] When the first determination condition is used to determine the current weather, if the first determination condition is met, it can be determined that the current weather is not bad weather. At this time, the current weather has been determined and no further determination is required. If the first determination condition is not met, it is impossible to determine whether the current weather is bad weather. If the total time period has not ended, the water signal can continue to be obtained. Similar to the above-mentioned embodiment, the bad weather to be determined may include but is not limited to at least one of the following: rain, hail, snow, fog, which have been explained and will not be repeated here.
[0320] Optionally, the first judgment condition may be a judgment condition set based on one or more specified parameters characterizing the water signal (for example, all or part of the shape of the water, the area covered by the water, the position covered by the water), which may be a judgment condition related to the water signal currently acquired, or it may be a judgment condition related to the water signal currently acquired and the water signal acquired previously or the time before that (in this case, the first judgment condition may be a judgment condition related to the difference between water signals acquired at different times).
[0321] Optionally, after the first determination condition is met, the current weather can continue to be determined based on the acquired water signal, or the process of determining the current weather based on the acquired water signal can be terminated, but the laser radar can still continue to detect the water signal on the shell surface of the laser radar, and after acquiring the water signal, the total time period can be reset and the above-mentioned current weather determination steps can be executed.
[0322] Through this embodiment, by determining whether the water stains on the surface of the laser radar are caused by bad weather based on the water signal obtained each time, and then determining the subsequent operations based on the determination result, the efficiency of current weather determination can be improved.
[0323] In an exemplary embodiment, the current weather is determined according to a first determination condition until the first determination condition is met or the total time period ends, including: when it is determined that the first determination condition is met based on the water signal obtained for the mth time, it is determined that the current weather is not bad weather; when none of the water signals obtained within the total time period meets the first determination condition, it is determined that the current weather is bad weather.
[0324] In this embodiment, the first determination condition may be one of the following: the water signal disappears; the water signal decreases to a specified value; the water signal decreases to a specified percentage, that is, if the water signal disappears, it can be determined that the current weather is not bad; if the water signal acquired after a period of time becomes significantly weaker, it can be determined that the current weather is not bad. Based on the above first determination condition, it can be determined that the water stains on the surface of the laser radar are not caused by bad weather.
[0325] Correspondingly, determining the current manner according to the first determination condition may include: determining that the current weather is not bad weather when the first determination condition is satisfied based on the water signal obtained for the mth time; and determining that the current weather is bad weather when the water signals obtained in the total time period do not satisfy the first determination condition. Here, m is a positive integer greater than or equal to 2.
[0326] Optionally, when determining the current weather, a first threshold value for obtaining water signals may be set, and the set first threshold value may be a positive integer greater than or equal to 2, and the first threshold value corresponds to the total time period. Correspondingly, water signals obtained by laser radar detecting the shell surface at least twice may be obtained until the first determination condition is satisfied based on a water signal obtained once, or the total time period ends (at this time, the number of times the water signal is obtained reaches the first threshold value).
[0327] For example, if the first threshold value is set to 3, the total time period can be understood as the time period from the moment when the water signal is first acquired to the moment when the water signal is third acquired (or a certain moment after the moment when the water signal is third acquired). In this case, when the number of acquired water signals reaches the first threshold value and the acquired water signals do not meet the first determination condition, it can be determined that the current weather is bad weather.
[0328] Through this embodiment, it is determined that the current weather is not bad weather based on the disappearance of the water signal or the reduction of the water signal to a specified value, so that the accuracy and efficiency of weather determination can be improved.
[0329] In an exemplary embodiment, a determination condition may be set for determining whether to continue acquiring a water signal, for example, a second determination condition, which may be a condition for determining that the water stains on the surface of the laser radar are not caused by bad weather. The second determination condition may be the same or similar to the first determination condition in the aforementioned embodiment, or may be a different determination condition. Correspondingly, the current weather is determined based on each acquired water signal, including: determining the current weather according to the second determination condition until the second determination condition is satisfied or the number of acquisitions of the water signal reaches a set threshold number of times.
[0330] The second number threshold defines the maximum number of times the water signal is acquired. The second number threshold and the first number threshold may be the same or different, may be pre-set, or may be adjusted according to user needs. For example, the second number threshold may be 3 or other values. If, before the number of times the water signal is acquired reaches the set second number threshold, it is determined that the second judgment condition is met based on the change of a water signal acquired once or two adjacent water signals acquired, then the current weather has been determined (i.e., it is not bad weather), and the above judgment process may be terminated; when the number of times the water signal is acquired reaches the set second number threshold, the water signal acquired once or any two adjacent water signals do not meet the second judgment condition, then the current weather has been determined (i.e., it is bad weather).
[0331] Optionally, when the water signal is acquired twice adjacently, the position change of the laser radar may be the same or different. For example, when the water signal is acquired twice adjacently, the distance moved by the laser radar or the lawn mower is the same.
[0332] Through this embodiment, the current weather judgment is performed based on the condition that the water stains on the surface of the laser radar are not caused by bad weather and the set number threshold, which can improve the efficiency of the current weather judgment.
[0333] In an exemplary embodiment, the current weather is determined according to a second determination condition until the second determination condition is met or the number of times the water signal is acquired reaches a set threshold number of times, including: when it is determined that the second determination condition is met based on changes in the water signal acquired for the nth time and the water signal acquired for the n-1th time, determining that the current weather is not bad weather; when the number of times the water signal is acquired reaches a set threshold number of times and the acquired water signals do not meet the second determination condition, determining that the current weather is bad weather.
[0334] In this embodiment, the second judgment condition may be: the coverage area of the water signal decreases and / or the coverage points of the water signal do not increase, that is, it can be determined that the water stains on the shell surface have not increased by the fact that the coverage area of the water signal obtained twice adjacently decreases and / or the coverage points of the water signal do not increase. It can be considered that the water stains on the shell surface have not increased due to rain, fog condensation and other phenomena caused by bad weather. It can be determined that the second judgment condition is not met, and it can be determined that the current weather is not bad weather.
[0335] Correspondingly, the method of determining the current weather according to the second determination condition may include: based on the change of the water signal obtained for the nth time and the water signal obtained for the n-1th time, when it is determined that the second determination condition is satisfied, determining that the current weather is not bad weather; when the number of times the water signal is obtained reaches a set number threshold, and the obtained water signals do not meet the second determination condition, determining that the current weather is bad weather. Here, n is a positive integer greater than or equal to 2. The second number threshold is similar to the first number threshold in the aforementioned embodiment, and has been explained, so it will not be repeated here.
[0336] Through this embodiment, the current weather is determined based on the changes in water signals obtained twice adjacently, which can improve the efficiency and accuracy of the current weather determination.
[0337] In an exemplary embodiment, the method further includes: determining the coverage area and coverage position of the acquired water signal according to the points where the water signal is detected on the shell surface, wherein whether the water signal coverage points are increased is determined based on the coverage position of the water signal.
[0338] In this embodiment, the coverage area of the water signal and the water signal coverage points can be determined by the points on the shell surface where the water signal is detected. A set of points can be configured for the shell surface of the laser radar (which can be pre-set or set after determining that the water signal needs to be acquired). When acquiring the water signal, it can be determined whether the water signal is detected at each configured point, and the points where the water signal is detected in the set of points can be obtained. The area covering the points on the shell surface where the water signal is detected can be used as the coverage area of the water signal, thereby determining the coverage area and coverage position of the acquired water signal.
[0339] When judging whether the water signal coverage points have increased, the judgment can be made directly based on the location and number of the points where the water signal is detected, or based on the coverage position of the water signal. For scenarios where the water signal coverage points are judged directly based on the points where the water signal is detected, if there is no other need to use the coverage area and coverage position of the water signal, the step of determining the coverage area and coverage position of the acquired water signal based on the points where the water signal is detected on the shell surface can be omitted.
[0340] Through this embodiment, the coverage area and coverage position of the acquired water signal are determined according to the point position of the water signal detected on the shell surface, which can improve the convenience of determining the coverage area and coverage position of the water signal.
[0341] The lawn mower control method in this embodiment is explained below in conjunction with an optional example. This optional example provides a solution for determining the current weather by whether there are water stains at different locations. When a water signal is detected for the first time, the point where water is detected on the sensor surface can be recorded, and the water coverage area and the corresponding coverage position can be calculated by an algorithm.
[0342] The lawn mower moves to the next position, records the points where water is detected on the sensor surface, and calculates the water coverage area and the corresponding coverage position through an algorithm. Compare the changes in water coverage between the two positions. If the water coverage area decreases and no new water coverage points are added, it indicates that there are no new water stains. Otherwise, enter the next position to continue detection and judgment, that is, continue to compare the changes in water coverage between the two positions. If the water coverage area decreases and no new water coverage points are added, it indicates that there are no new water stains. Similarly, when entering the second to last position, continue detection and judgment, that is, continue to compare the changes in water coverage between the two positions. If the water coverage area decreases and no new water coverage points are added, it indicates that there are no new water stains. Otherwise, it is judged that there are new water stains, and then the current weather is judged to be bad weather, and the lawn mower is controlled to perform corresponding operations.
[0343] It should be noted that each time the lawn mower enters the next position, the direction of the laser radar can be controlled to be different, or the distance between each two adjacent positions can be a specified distance, and the next position entered by the lawn mower can also be limited by other methods.
[0344] Through this optional example, the accuracy of weather determination can be improved by controlling the lawn mower to detect water signals at different positions and judging the current weather based on the changes in the water signals detected between two adjacent times.
[0345] In an exemplary embodiment, when the current weather is judged to be severe weather, a shielding action is performed on at least a portion of the lawn mower, including: when the current weather is judged to be severe weather, controlling the lawn mower to return to a base station matched with the lawn mower.
[0346] In this embodiment, if the current weather is judged to be bad weather, in order to prevent the bad weather from damaging the lawn mower, the lawn mower can be controlled to return to the base station that matches the lawn mower. The location of the base station in this embodiment and the setting method of the matching relationship between the base station and the lawn mower are similar to those in the previous embodiment, and have been explained, so they will not be repeated here.
[0347] Through this embodiment, when it is determined that the current weather is bad weather, the lawn mower is controlled to return to the base station that matches the lawn mower, which can improve the safety of the lawn mower operation and improve the timeliness of the lawn mower control.
[0348] In an exemplary embodiment, when the current weather is judged to be bad weather, shielding at least part of the lawn mower includes: when the current weather is judged to be bad weather, controlling the lawn mower to move to a designated position so that at least part of the lawn mower is shielded. The designated position referred to in this embodiment can be under the eaves of a building, under an awning, or any sheltered place. After the lawn mower moves to the designated position, it can continue to perform the mowing action at the designated position, or it can stop the mowing action and stop somewhere.
[0349] In this embodiment, if the current weather is judged to be bad weather, the lawn mower can be controlled to move to a designated location to avoid damage to the lawn mower caused by the bad weather. The designated location and the setting method of the designated location in this embodiment are similar to those in the previous embodiment, and have been explained, so they will not be repeated here.
[0350] In an exemplary embodiment, when the current weather is judged to be severe weather, a shielding action is performed on at least part of the lawn mower, including: when the current weather is judged to be severe weather, a shielding member is controlled to shield the laser radar, and the shielding member may be arranged on the lawn mower or outside the lawn mower. When the shielding member is arranged on the lawn mower, a control unit may be used to control the shielding member to shield the laser radar under severe weather conditions, and to be retracted under non-severe weather conditions to expose the laser radar so that the lawn mower can work normally; when the shielding member is arranged outside the lawn mower, a control unit may be used to control the shielding member to move to the position of the laser radar under severe weather conditions to shield the laser radar, and to leave the laser radar under non-severe weather conditions to remove the shielding of the laser radar.
[0351] In an exemplary embodiment, when the current weather is judged to be bad weather, a shielding action is performed on at least part of the lawn mower, including: when the current weather is judged to be bad weather, a shielding member is controlled to shield the body of the laser radar, and the shielding member can be set on the lawn mower or outside the lawn mower. When the shielding member is set on the lawn mower, a control unit can be used to unfold the shielding member to form a shield for the laser radar or retract the shielding member to remove the shielding for the laser radar; when the shielding member is set outside the lawn mower, the shielding member can be a movable shielding shed. When it is judged that the current weather is bad weather, the shielding shed can be moved to the outside of the lawn mower to form a shield for the lawn mower. When it is judged that the current weather is not bad weather, the shielding shed can leave the lawn mower so that the lawn mower can work normally.
[0352] Through this embodiment, when it is determined that the current weather is inclement weather, the lawn mower is controlled to move to a designated position, which can improve the safety of the lawn mower operation and increase the service life of the lawn mower.
[0353] In an exemplary embodiment, the method further comprises: in the case where there is water stain on the surface of the shell, controlling the lawn mower to stop mowing, wherein the cutting assembly of the lawn mower is lifted after the mowing is stopped.
[0354] The lawn mower further comprises a cutting assembly for performing the mowing operation. The position, working mode, type, etc. of the cutting assembly are similar to those in the above-mentioned embodiment and have been described, so they will not be described in detail here.
[0355] If there are water stains on the shell surface, considering that the water stains may be caused by bad weather, in order to ensure the safe operation of the cutting component, when there are water stains on the shell surface, the lawn mower is controlled to stop mowing. After stopping the mowing work, the cutting component of the lawn mower is lifted up, so as to improve the convenience of moving the lawn mower and increase the service life of the cutting component.
[0356] Through this embodiment, when there are water stains on the shell surface, the lawn mower can be controlled to stop mowing. After stopping mowing, the cutting component of the lawn mower is lifted up, which can reduce the risk of damage to the cutting component, improve the convenience of moving the lawn mower, and thus increase the service life of the cutting component.
[0357] In an exemplary embodiment, the lawn mower is controlled to perform corresponding actions according to the current weather, including: when it is determined that the current weather is not severe weather, the lawn mower is controlled to continue mowing according to a preset cutting path and cutting mode.
[0358] In this embodiment, in order to ensure the efficiency of mowing, when it is determined that the current weather is not bad weather, the mower can be controlled to continue mowing, for example, to continue mowing according to a pre-set cutting path and cutting mode. The method of controlling the mower to continue mowing is similar to that in the above-mentioned embodiment, which has been explained and will not be repeated here.
[0359] Through this embodiment, when it is determined that the current weather is not bad weather, the lawn mower is controlled to continue mowing according to the pre-set cutting path and cutting mode, thereby ensuring the mowing efficiency of the lawn mower.
[0360] In an exemplary embodiment, the method further includes: when there is water stain on the shell surface, before determining the current weather, the lawn mower continues mowing according to a preset cutting path and cutting mode.
[0361] In this embodiment, in order to ensure the efficiency of mowing, before the current weather is determined, the mower can be controlled to continue mowing, for example, to continue mowing according to a preset cutting path and cutting mode. The method of controlling the mower to continue mowing is similar to that in the above embodiment, which has been described and will not be repeated here.
[0362] Through this embodiment, before the current weather is determined, the lawn mower is controlled to continue mowing according to a preset cutting path and cutting mode, thereby ensuring the mowing efficiency of the lawn mower.
[0363] According to another aspect of the embodiment of the present application, a lawn mower control method is also provided. The method embodiment provided in the embodiment of the present application can be used in a lawn mower (such as Figure 1 As shown), or a similar processing device is executed, which has been explained and will not be repeated here.
[0364] The lawn mower control method provided in this embodiment can be executed by the lawn mower alone, or can be executed by the user end or the base station and the lawn mower together. Fig.11 As shown, the process of the lawn mower control method includes the following steps:
[0365] S1102, when there is water stain on the shell surface of the laser radar, obtaining a water signal obtained by the laser radar detecting the shell surface at least once;
[0366] S1104: When the coverage value of the single water signal increased on the shell surface per unit time is greater than a preset value, the lawn mower is controlled to continue mowing according to a preset cutting path and cutting mode.
[0367] Similar to the aforementioned embodiment, the lawn mower control method in this embodiment can be applied to the scenario of target detection for a lawn mower or similar equipment. The lawn mower may include a laser radar, the number and setting position of the laser radar, the characterization parameters of the water signal, and the way in which the lawn mower obtains the water signal by detecting the shell surface by the laser radar is similar to the aforementioned embodiment and will not be repeated here. During the operation of the lawn mower, the laser radar can detect the shell surface and obtain the laser radar data; by parsing the laser radar data, the water signal of the above-mentioned shell surface can be obtained. In the case where there are water stains on the shell surface of the laser radar, the water signal obtained by the laser radar detecting the shell surface once can be obtained.
[0368] Considering that the water stains on the surface of the lawn mower are not necessarily caused by bad weather (for example, rain, fog condensation, hail melting, etc.), there may be other sources, for example, there is a possibility that water signals are detected on the surface of the lawn mower due to human splashing or other machines spraying water. For water stains caused by bad weather, it takes a certain amount of time to cover the shell surface, while for water stains caused by reasons such as human splashing, the water stains will cover the shell surface in large quantities and quickly; in addition, water stains caused by bad weather are more evenly covered than water stains caused by other reasons, and water stains caused by other reasons usually cover local areas on the shell surface. Therefore, whether the generation of the water signal is caused by bad weather is determined based on whether the coverage value of a single water signal added to the shell surface per unit time is greater than a preset value: when it is detected that the coverage value of a single water signal added to the shell surface per unit time is greater than a preset value, it is determined that the source of the water signal may be other reasons other than bad weather, and the current weather is not bad weather. Exemplarily, bad weather includes at least one of rain, hail, snow, and fog.
[0369] Here, the specific duration of the unit time and the preset value of the water signal can be pre-set and can be adjusted according to user needs. The value can be determined based on analysis of historical data (for example, based on experience value setting) or can be determined based on statistical experimental results. For example, the unit time can be 1s, 2s or other values, the preset value can be a preset water signal coverage area, or it can be other parameters, which are not limited in this embodiment.
[0370] Here, a single water signal can be used to represent a single continuous and closed water stain area, that is, it can be determined whether the water signal represents a single continuous and closed water stain area based on one or more specified parameters characterized by the water signal (for example, the water coverage area, the water coverage position, the water shape, etc.), and the judgment and analysis process of the water signal characterization parameters is similar to that in the aforementioned embodiment. The coverage value of a single water signal can be calculated based on one or more specified parameters characterized by the water signal, for example, it can be calculated based on the water signal area and the water signal strength, and the calculation process and the use parameters can be pre-set or adjusted according to user needs.
[0371] According to the current weather, the lawn mower can be controlled to perform corresponding actions. For example, the lawn mower can be controlled to continue working or return to the corresponding base station according to whether the current weather is bad weather, or the lawn mower can be controlled to continue working or perform other preset actions according to the specific weather type of the current weather. In addition, the user can also control the lawn mower to perform the corresponding action selected by the user through an interactive interface or a control terminal. When it is detected that the coverage value of a single water signal added to the shell surface per unit time is greater than a preset value, it can be determined that the source of the water signal may be other reasons other than bad weather, that is, it can be determined that the current weather is not bad weather, and then the lawn mower can be controlled to perform corresponding actions, including: controlling the lawn mower to continue mowing according to a pre-set cutting path and cutting mode, thereby ensuring the working efficiency of the lawn mower.
[0372] Through the above steps, when there are water stains on the shell surface of the laser radar, a water signal obtained by the laser radar detecting the shell surface once is obtained; when the coverage value of a single water signal increased on the shell surface per unit time is greater than a preset value, it is determined that the current weather is not bad weather; and the lawn mower is controlled to perform corresponding actions according to the current weather, which can eliminate the situation where water signals are detected on the shell surface due to reasons other than bad weather, thereby improving the accuracy of the current weather judgment.
[0373] In an exemplary embodiment, the coverage value of a single water signal added to the shell surface per unit time is greater than a preset value, including: the coverage area of a single water signal added to the shell surface per unit time is greater than or equal to one third of the shell surface area.
[0374] Considering that water stains caused by severe weather usually cover the shell surface relatively slowly and evenly, and will not cover the shell surface in a short time, a threshold (an optional example of a preset value) of the coverage area of a single water signal added per unit time can be set, for example, one-third or one-half of the shell surface area. If it is detected that the coverage area of a single water signal added per unit time on the shell surface reaches the set coverage area threshold (for example, one-third or one-half of the shell surface area), it can be determined that the current weather is not severe weather.
[0375] Through this embodiment, by comparing the coverage area of a single water signal added to the shell surface per unit time with a set threshold of the coverage area, it is determined whether the water stains on the shell surface are caused by bad weather, thereby improving the accuracy of current weather judgment.
[0376] In an exemplary embodiment, the coverage value of a single water signal added to the shell surface per unit time is greater than a preset value, including: the coverage area of a single water signal added to the shell surface per unit time is greater than or equal to 6 square centimeters.
[0377] In this embodiment, considering that the size differences of the laser radars on lawn mowers are usually not large, in order to improve the convenience of current weather judgment, the above-mentioned coverage area threshold (specified value) can be a specified area threshold, which has a faster calculation speed than the aforementioned coverage area threshold set according to a ratio (for example, one-third) because there is no need to know the area of the shell surface and no need to calculate the ratio.
[0378] In this embodiment, the specified area threshold may be 6 square centimeters. When the coverage area of a single water signal added to the shell surface per unit time is greater than or equal to 6 square centimeters, it can be determined that the current weather is not severe weather. The specified area threshold may also be other values. Compared with other area thresholds, 6 square centimeters can adapt to actual scenarios and improve the accuracy of current weather determination.
[0379] Through this embodiment, when the coverage area of a single water signal added to the shell surface per unit time is greater than or equal to 6 square centimeters, it is determined that the current weather is not bad weather, which can improve the accuracy of the current weather determination.
[0380] In an exemplary embodiment, the coverage value of a single water signal added to the shell surface per unit time is greater than a preset value, including: the diameter of a single water signal added to the shell surface per unit time is greater than or equal to 2 centimeters.
[0381] Considering that water stains generated by severe weather are usually relatively slow and usually not too large, a threshold value (an optional example of a preset value) of the diameter of a single water signal increased per unit time can be set, for example, 2 cm, 3 cm, 5 cm, etc. If the diameter of a single water signal increased on the shell surface per unit time is greater than or equal to the set diameter threshold value (for example, 2 cm), it can be determined that the current weather is not severe weather.
[0382] Through this embodiment, by comparing the diameter of a single water signal increased on the shell surface per unit time with a set diameter threshold, it is determined whether the water stains on the shell surface are caused by bad weather, thereby improving the accuracy of current weather judgment.
[0383] In an exemplary embodiment, the length of the unit time affects not only the timeliness of the current weather determination, but also the accuracy of the detection. For example, if the length of the unit time is too short, it may not be possible to determine the bad weather. If the length of the unit time is too long, it may not be possible to determine the bad weather in time, increasing the risk of bad weather causing damage to the lawn mower. It may also lead to inaccurate current weather determination due to the inability to accurately characterize the current weather due to the change of the water signal. In this regard, in order to improve the timeliness and accuracy of the current weather determination, the length of the unit time is less than or equal to 2 seconds.
[0384] Through this embodiment, by setting the length of the unit time to be less than or equal to 2 seconds, the timeliness and accuracy of the current weather determination can be improved.
[0385] Obviously, those skilled in the art should understand that the modules or steps of the above-mentioned embodiments of the present application can be implemented by a general computing device, they can be concentrated on a single computing device, or distributed on a network composed of multiple computing devices, they can be implemented by a program code executable by a computing device, so that they can be stored in a storage device and executed by the computing device, and in some cases, the steps shown or described can be executed in a different order from that herein, or they can be made into individual integrated circuit modules, or multiple modules or steps therein can be made into a single integrated circuit module for implementation. In this way, the embodiments of the present application are not limited to any specific combination of hardware and software.
[0386] The above are only preferred embodiments of the present application and are not intended to limit the embodiments of the present application. For those skilled in the art, the embodiments of the present application may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the principles of the embodiments of the present application shall be included in the protection scope of the embodiments of the present application.
Claims
1. A target detection method for a lawn mower, characterized in that: include: Acquire a target frame set obtained by detecting a sensor shell surface of the target sensor through a target sensor, wherein each target frame in the target frame set is a single-frame point cloud obtained by one detection by the target sensor, and the target sensor is a sensor configured on the lawn mower for generating a point cloud; Performing a projection operation on the target frame set to obtain a target projection image, wherein the projection operation is an operation of projecting onto a specified plane; The target projection image is input into a trained target classification model to obtain a target classification result, wherein the target classification model is a classification model used to distinguish multiple preset categories, and the target classification result is used to indicate a preset category among the multiple preset categories that matches the target frame set.
2. The method according to claim 1, characterized in that The acquiring of a target frame set obtained by detecting a sensor shell surface of the target sensor through the target sensor includes: Acquire multiple groups of echo signals obtained by transmitting detection signals multiple times by the target sensor to the surface of the sensor housing, wherein each group of echo signals in the multiple groups of echo signals includes two echo signals corresponding to a detection signal transmitted once; The two echo signals in each group of echo signals are fused into a corresponding single-frame point cloud to obtain the target frame set.
3. The method according to claim 2, characterized in that One echo signal in each group of echo signals is an echo signal reflected by the surface of the sensor housing, and the other echo signal is an echo signal reflected by the detected target.
4. The method according to claim 2, characterized in that: The step of fusing two echo signals in each group of echo signals into a corresponding single-frame point cloud to obtain the target frame set includes: Based on the signal difference at the same position point of two echo signals in each group of echo signals, a single frame point cloud corresponding to each group of echo signals is determined to obtain the target frame set, wherein in the target frame set, the value of a position point corresponds to the signal difference between the two echo signal position points corresponding to the one position point.
5. The method according to claim 1, characterized in that Each position point in each target frame is represented by a set of position parameters, and the value of each position point in each target frame is used to represent the degree of occlusion of the corresponding position point; Before performing the projection operation on the target frame set, the method further includes: A filtering operation is performed on the target frame set based on at least one position parameter to obtain the filtered target frame set, wherein the at least one position parameter belongs to the group of position parameters.
6. The method according to claim 5, characterized in that The performing a filtering operation on the target frame set based on at least one position parameter to obtain the filtered target frame set includes at least one of the following: Filter out the position points in the target frame set whose corresponding modulus length is not a specified value, and obtain the target frame set after modulus length filtering; Position points in the target frame set whose corresponding polar angles are greater than or equal to a specified angle are filtered out to obtain the target frame set after polar angle filtering.
7. The method according to claim 5, characterized in that The set of position parameters includes three axis coordinates in a rectangular coordinate system; the tilt angle of the target sensor includes a roll angle and a pitch angle; The performing a filtering operation on the target frame set based on at least one position parameter to obtain the filtered target frame set includes: Using a first rotation matrix and a second rotation matrix to transform the coordinate values of the three axis coordinates of each position point in the target frame set, to obtain the target frame set after coordinate transformation, wherein the first rotation matrix is a rotation matrix determined according to the roll angle, and the second rotation matrix is a rotation matrix determined according to the pitch angle; The position points whose corresponding z-axis coordinates are less than a preset value in the target frame set after coordinate transformation are filtered out to obtain the target frame set after coordinate filtering.
8. The method according to claim 1, characterized in that The performing a projection operation on the target frame set to obtain a target projection image includes: Projecting the target frame set into the specified plane to obtain target projection information corresponding to the target frame set, wherein the target projection information is used to represent the pixel position to which each position point in the target frame set is projected and the pixel value corresponding to each position point, and the pixel value corresponding to each position point is the pixel value of the pixel point projected into the specified plane by each position point; The target projection image is generated according to the pixel position to which each position point is projected and the pixel value corresponding to each position point.
9. The method according to claim 8, characterized in that The step of generating the target projection image according to the pixel position to which each position point is projected comprises: In the case where there is a position point set projected to the same pixel position in the target frame set, an average value of pixel values corresponding to the position points in the position point set is determined as the pixel value of the pixel position projected to the position point set in the target projection image; Determine the pixel values corresponding to other position points in the target frame set except the position points in the position point set as the pixel values of the pixel positions to which the other position points in the target projection image are projected; When there is a group of pixel positions whose pixel values are not determined within the image area of the target projection image, the designated pixel value is determined as the pixel value of the group of pixel positions in the target projection image.
10. The method according to claim 1, characterized in that The target frame set is obtained by the target sensor detecting the surface of the sensor shell in one of a plurality of continuous time intervals; the classification result corresponding to each time interval in the plurality of continuous time intervals is the confidence level output by the target classification model and corresponding to each preset category in the plurality of preset categories; After inputting the target projection image into a trained target classification model to obtain a target classification result, the method further includes: Smoothing multiple confidences corresponding to the same preset category in the classification results corresponding to each time interval to obtain a smoothed confidence corresponding to each preset category; The preset category with the highest corresponding smoothing confidence among the multiple preset categories is determined as the target category corresponding to the multiple continuous time intervals.
11. The method according to claim 1, characterized in that: The method further comprises: Performing a training sample acquisition operation multiple times based on a reference sensor to obtain a training image sample set, wherein the reference sensor and the target sensor are sensors of the same model, and the labeling information of each training image sample in the training image sample set is used to indicate one of the multiple preset categories; Performing model training on the target classification model to be trained based on each training image sample and the annotation information of each training image sample to obtain the trained target classification model; The training sample collection operation includes: Acquire a reference frame set obtained by detecting the sensor shell surface of the reference sensor through the reference sensor, wherein each reference frame in the reference frame set is a single frame point cloud obtained by performing one detection by the reference sensor; A projection operation is performed on the reference frame set to obtain a training image sample.
12. The method according to any one of claims 1 to 11, characterized in that The acquiring of a target frame set obtained by detecting a sensor shell surface of the target sensor through the target sensor includes: The target frame set obtained by detecting the laser radar shell surface of the multi-dimensional laser radar through the multi-dimensional laser radar is obtained, wherein the target sensor is the multi-dimensional laser radar.
13. The method according to claim 12, characterized in that The multi-dimensional laser radar is a three-dimensional laser radar.
14. The method according to any one of claims 1 to 11, characterized in that The target sensor is an integrated sensing device, and the target sensor is also used for at least one of the following: navigation, positioning, and obstacle avoidance.
15. The method according to any one of claims 1 to 11, characterized in that The plurality of preset categories are used to represent at least one of the following targets: rain, hail, snow, sandstorm, and fog.
16. The method according to any one of claims 1 to 11, characterized in that The surface of the sensor housing is a plane and / or a curved surface.
17. The method according to any one of claims 1 to 11, characterized in that The target sensor is disposed at a front end region of the top of the lawn mower at an angle to the horizontal plane.
18. A target detection device for a lawn mower, characterized in that: include: an acquisition unit, used for acquiring a target frame set obtained by detecting a sensor shell surface of a target sensor through a target sensor, wherein each target frame in the target frame set is a single frame point cloud obtained by a detection performed by the target sensor, and the target sensor is a sensor configured on the lawn mower for generating a point cloud; A first execution unit is used to perform a projection operation on a target frame set to obtain a target projection image, wherein the projection operation is an operation of projecting onto a specified plane; The input unit is used to input the target projection image into the trained target classification model to obtain a target classification result, wherein the target classification model is a classification model used to distinguish multiple preset categories, and the target classification result is used to indicate a preset category among the multiple preset categories that matches the target frame set.
19. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored program, wherein the program executes the method according to any one of claims 1 to 17 when executed.
20. An electronic device comprising a memory and a processor, characterized in that: A computer program is stored in the memory, and the processor is configured to execute the method according to any one of claims 1 to 17 through the computer program.
21. A lawn mower control method, characterized in that: The lawn mower includes a laser radar, and the method includes: In the case where there is water stain on the shell surface of the laser radar, obtaining a water signal obtained by the laser radar detecting the shell surface at least once, and judging the current weather based on the obtained water signal; Controlling the lawn mower to perform corresponding actions according to the current weather; The controlling the lawn mower to perform corresponding actions according to the current weather includes: when it is determined that the current weather is bad weather, shielding at least part of the lawn mower.
22. The method according to claim 21, characterized in that Before determining the current weather, the lawn mower continues mowing according to the preset cutting path and cutting mode.
23. The method according to claim 21, characterized in that The shielding action performed on at least part of the lawn mower comprises: The lawn mower is controlled to return to a base station matched with the lawn mower, or the lawn mower is controlled to move to a designated position so that at least a part of the lawn mower is shielded.
24. The method according to claim 21, characterized in that The shielding action performed on at least part of the lawn mower comprises: A shielding member is controlled to shield at least a portion of the lawn mower to isolate at least a portion of the lawn mower from external severe weather.
25. The method according to claim 21, characterized in that The controlling the lawn mower to perform corresponding actions according to the current weather also includes: When it is determined that the current weather is bad weather, the lawn mower is controlled to stop mowing, wherein the cutting component of the lawn mower is lifted after the mowing is stopped.
26. The method according to claim 21, characterized in that The controlling the lawn mower to perform corresponding actions according to the current weather also includes: When it is determined that the current weather is not severe weather, the lawn mower is controlled to continue mowing according to a preset cutting path and cutting mode.
27. The method according to claim 26, characterized in that In the case where there is water stain on the shell surface of the laser radar, obtaining a water signal obtained by the laser radar detecting the shell surface at least once, and judging the current weather based on the obtained water signal, including: A target frame set of water signals obtained by the laser radar detecting the shell surface at least once is obtained, the target frame set is processed and compared with a preset classification model, and the current weather is determined according to the target classification result, wherein the preset classification model is classified according to different types of weather.
28. The method according to claim 27, characterized in that The step of obtaining a target frame set of water signals obtained by the laser radar detecting the shell surface at least once, processing the target frame set and comparing it with a preset classification model, and determining the current weather according to the target classification result includes: The shell surface of the laser radar is detected J times within a detection cycle. If the judgment results of the first K detections in the J detections are all that the current weather is bad weather, then the final weather output is bad weather; if the judgment results of the first L detections in the J detections are all that the current weather is bad weather, and the judgment result of the L+1th detection is that the current weather is not bad weather, then enter the next detection cycle until the current weather is determined; if at least one detection result among the judgment results of the first L detections in the J detections is that the current weather is not bad weather, then the final output is that the current weather is not bad weather, where L<K≤J and J≥2.
29. The method according to claim 28, characterized in that The J=3, the K=3, and the L=2.
30. The method according to claim 28, characterized in that The duration of the detection cycle is negatively correlated with the temperature of the shell surface.
31. The method according to claim 30, characterized in that The duration of one detection cycle is 10-30S.
32. The method according to any one of claims 21 to 31, characterized in that The severe weather includes at least one of rain, hail, snow and fog.
33. A lawn mower control method, characterized in that: The lawn mower includes a laser radar, and the method includes: In the case where there is water stain on the shell surface of the laser radar, a water signal obtained by the laser radar detecting the shell surface at least twice within a set total time period, and the current weather is determined based on the obtained water signal; Controlling the lawn mower to perform corresponding actions according to the current weather; The controlling the lawn mower to perform corresponding actions according to the current weather includes: when it is determined that the current weather is bad weather, shielding at least part of the lawn mower.
34. The method according to claim 33, characterized in that The shielding action performed on at least part of the lawn mower comprises: The lawn mower is controlled to return to a base station matched with the lawn mower; or the lawn mower is controlled to move to a designated position so that at least a part of the lawn mower is shielded.
35. The method according to claim 33, characterized in that The shielding action performed on at least part of the lawn mower comprises: A shielding member is controlled to shield at least a portion of the lawn mower to isolate at least a portion of the lawn mower from external severe weather.
36. The method according to claim 33, characterized in that The controlling the lawn mower to perform corresponding actions according to the current weather also includes: When it is determined that the current weather is bad weather, the lawn mower is controlled to stop mowing, wherein the cutting component of the lawn mower is lifted after the mowing is stopped.
37. The method according to claim 33, characterized in that The controlling the lawn mower to perform corresponding actions according to the current weather also includes: When it is determined that the current weather is not severe weather, the lawn mower is controlled to continue mowing according to a preset cutting path and cutting mode.
38. The method according to claim 33, characterized in that Before determining the current weather, the lawn mower continues mowing according to the preset cutting path and cutting mode.
39. The method according to claim 33, characterized in that Before acquiring at least twice the water signal obtained by the laser radar detecting the shell surface within the set total time period, the method further includes: The duration of the total time period is set according to the temperature of the shell surface, wherein the duration of the total time period is negatively correlated with the temperature of the shell surface.
40. The method according to claim 39, characterized in that When the temperature of the shell surface is 35-60° C., the total time period is 10-30 seconds.
41. The method according to claim 33, characterized in that The total time period is divided into a plurality of sub-time periods; The step of acquiring the water signal obtained by detecting the shell surface by the laser radar at least twice within the set total time period includes: Within the set total time period, a water signal obtained by the laser radar detecting the shell surface is acquired once at the end of each sub-time period in the multiple sub-time periods.
42. The method according to claim 41, characterized in that The durations of the sub-time periods within the total time period are equal.
43. The method according to claim 41, characterized in that When the total time period is 15 seconds, the total time period is divided into three sub-time periods, and the length of each sub-time period is 5 seconds.
44. The method according to claim 33, characterized in that The step of acquiring the water signal obtained by detecting the shell surface by the laser radar at least twice within the set total time period includes: Within a set total time period, water signals obtained by the laser radar detecting the shell surface at different positions are obtained.
45. The method according to claim 44, characterized in that The step of obtaining a water signal obtained by detecting the shell surface by the laser radar at different positions includes: A water signal obtained by the laser radar detecting the shell surface when the lawn mower moves to different positions is obtained, wherein the posture of the laser radar is different when the lawn mower moves to different positions.
46. The method according to claim 44, characterized in that The step of obtaining a water signal obtained by detecting the shell surface by the laser radar at different positions includes: When the lawn mower is in working state, obtaining a water signal obtained by the laser radar detecting the shell surface for the first time; Control the lawn mower to move or rotate to a detection posture that meets the signal detection conditions, and obtain a water signal obtained by the laser radar detecting the shell surface when the lawn mower is in the detection posture that meets the signal detection conditions, wherein the signal detection conditions include at least one of the following: a direction different from that when the water signal was obtained last time; a position different from that when the water signal was obtained last time; and a distance between the position and the position when the water signal was obtained last time is a specified distance.
47. The method according to claim 33, characterized in that The determining of the current weather based on the acquired water signal comprises: The current weather is determined according to a first determination condition until the first determination condition is satisfied or the total time period ends, wherein the first determination condition is a condition for determining that the water stains on the surface of the laser radar are not caused by bad weather.
48. The method according to claim 47, characterized in that The determining the current weather according to the first determination condition until the first determination condition is satisfied or the total time period ends includes: When it is determined that the first determination condition is satisfied based on the water signal obtained for the mth time, it is determined that the current weather is not bad weather, where m is a positive integer greater than or equal to 2; If the water signals acquired within the total time period do not satisfy the first determination condition, determining that the current weather is bad weather; The first determination condition is one of the following: the water signal disappears; the water signal decreases to a specified value relative to the first acquired water signal; the water signal decreases to a specified percentage compared to the first acquired water signal.
49. The method according to claim 33, characterized in that The determining of the current weather based on the acquired water signal comprises: The current weather is determined according to a second determination condition until the second determination condition is met or the number of acquisitions of the water signal reaches a set threshold number of times, wherein the second determination condition is a condition for determining that the water stains on the surface of the laser radar are not caused by bad weather.
50. The method according to claim 49, characterized in that The determining the current weather according to the second determination condition until the second determination condition is satisfied or the number of acquisitions of the water signal reaches a set number threshold includes: When it is determined that the second determination condition is satisfied based on the change of the water signal obtained for the nth time and the water signal obtained for the (n-1th time), it is determined that the current weather is not raining, where n is a positive integer greater than or equal to 2; When the number of acquisitions of the water signal reaches a set number threshold and the acquired water signals do not meet the second determination condition, determining that the current weather is bad weather; Wherein, the second determination condition is: the coverage area of the water signal decreases and / or the coverage points of the water signal do not increase.
51. The method according to claim 50, characterized in that The method further comprises: The coverage area and coverage position of the acquired water signal are determined according to the points where the water signal is detected on the shell surface, wherein whether the water signal coverage points are increased is determined based on the coverage position of the water signal.
52. The method according to any one of claims 33 to 51, characterized in that The severe weather includes at least one of rain, hail, snow and fog.
53. A lawn mower control method, characterized in that: The lawn mower includes a laser radar, and the method includes: In the case where there is water stain on the shell surface of the laser radar, obtaining a water signal obtained by the laser radar detecting the shell surface at least once; When the coverage value of the single water signal increased on the shell surface per unit time is greater than a preset value, the lawn mower is controlled to continue mowing according to a preset cutting path and cutting mode.
54. The method according to claim 53, characterized in that The coverage value of a single water signal added to the shell surface per unit time is greater than a preset value, including: The coverage area of a single water signal added to the shell surface per unit time is greater than or equal to one third of the shell surface area.
55. The method according to claim 53, characterized in that The coverage value of a single water signal added to the shell surface per unit time is greater than a preset value, including: The coverage area of a single water signal added to the shell surface per unit time is greater than or equal to 6 square centimeters.
56. The method according to claim 53, characterized in that The coverage value of a single water signal added to the shell surface per unit time is greater than a preset value, including: The diameter of a single water signal added on the shell surface per unit time is greater than or equal to 2 centimeters.
57. The method according to claim 53, characterized in that The length of the unit time is less than or equal to 2 seconds.
Citation Information
Patent Citations
Methods and systems for detecting weather conditions using vehicle onboard sensors
CN105324287A
Barrier identification method and apparatus, computer device and readable medium
CN106934347A
Three-dimensional point cloud segmentation method and device and movable platform
CN114556419A
Mower charging system and charging base station thereof
CN210693492U
Cited By
Mower ground type identification algorithm based on millimeter wave radar and application method thereof
CN120507727A
Control method for self-moving device, and electronic device
WO2026219271A1
