Control Method, Device and Storage Medium of Self-Moving Device

Abstract

The present application relates to a control method, device and storage medium for a self-moving device, belonging to the field of computer technology. The method includes: obtaining edge information of the working area where the self-moving device is located; obtaining environmental images collected during movement; obtaining access door information based on the environmental images; dividing independent areas within the working area based on the access door information and the edge information; which can solve the problem that the prior art cannot divide the working area; and can divide the working area to achieve the effect of personalized work according to the divided independent areas. In addition, since the access door can be an open virtual door, therefore, the information of the virtual door and the edge information can be combined to obtain the regional boundary of each independent area, so as to divide the corresponding independent area, and the regional division of the open door scenario can be realized, improving the accuracy of regional division.

Description

Technical Field

[0001] This application relates to a control method, device, and storage medium for a self - moving device, belonging to the field of computer technology. Background Art

[0002] With the development of artificial intelligence and the robotics industry, intelligent household appliances such as floor - cleaning robots have gradually become popular.

[0003] Common floor - cleaning robots collect environmental pictures through a camera component fixed above the machine body. Using image recognition algorithms, they can only identify the items in the collected pictures and cannot perform environmental scene recognition. Summary of the Invention

[0004] This application provides a control method, device, and storage medium for a self - moving device, which can solve the problem that the prior art cannot divide the working area of the self - moving device. The following technical solutions are provided in this application:

[0005] In a first aspect, a control method for a self - moving device is provided. The method includes:

[0006] Obtain the edge information of the working area where the self - moving device is located;

[0007] Obtain the environmental image collected during movement;

[0008] Obtain the access door information based on the environmental image;

[0009] Based on the access door information and the edge information, divide the independent areas within the working area.

[0010] Optionally, after dividing the independent areas within the working area based on the access door information and the edge information, the method further includes:

[0011] Determine the scene prediction result of the corresponding independent area based on the environmental image;

[0012] Determine the scene type of the independent area according to the scene prediction result of the independent area.

[0013] Optionally, the step of dividing the independent areas within the working area based on the access door information and the edge information includes:

[0014] Obtain the position information of the corresponding access door indicated by the access door information in the working area;

[0015] Combine the edge information and the position information to obtain the combined boundary information;

[0016] Divide each closed area formed by the combined boundary information into corresponding independent areas.

[0017] Optionally, obtaining the access door information based on the environmental image includes:

[0018] Identify whether the environmental image includes an image of an access door;

[0019] When the environmental image includes an image of an access door, obtain the position information of the access door in the working area.

[0020] Optionally, the access door includes the door frame in the working area.

[0021] Optionally, determining the scene type of the independent area according to the scene prediction result of the independent area includes:

[0022] Obtain the pose information of each independent area, where the pose information includes the position information and direction information of the corresponding independent area in the working area;

[0023] Combine the scene area result of each independent area and the pose information of each independent area, and determine the scene type of each independent area according to a preset probability distribution strategy;

[0024] Among them, the probability distribution strategy is used to determine, for each target scene type, the independent area with the highest probability of having the scene type as the target scene type from each independent area.

[0025] Optionally, determining the scene prediction result of the corresponding independent area based on the environmental image includes:

[0026] Obtain an image recognition model, where the computing resources occupied when the image recognition model runs are lower than the maximum computing resources provided by the self - moving device;

[0027] For each independent area, input the environmental image corresponding to the independent area into the image recognition model to obtain an object recognition result, where the object recognition result includes the attribute information of the target object;

[0028] Obtain a scene recognition model, where the scene recognition model is trained using sample attribute information of objects and sample scene types;

[0029] Input the object recognition result into the scene recognition model to obtain a scene prediction result, where the scene prediction result includes at least one predicted scene type of the independent area.

[0030] Optionally, the image recognition model is trained based on a small network model using sample environmental images and sample object results.

[0031] Optionally, the scene recognition model is trained based on a probability model using the sample attribute information and sample scenes of the object.

[0032] In a second aspect, a control device for a self - moving device is provided. The device includes:

[0033] A first information acquisition module, configured to acquire the edge information of the working area where the self - moving device is located;

[0034] An environmental image acquisition module, configured to acquire the environmental images collected during movement;

[0035] A second information acquisition module, configured to obtain the access door information based on the environmental images;

[0036] A region division control module, configured to divide the independent regions within the working area based on the access door information and the edge information.

[0037] In a third aspect, a control device for a self - moving device is provided. The device includes a processor and a memory; a program is stored in the memory, and the program is loaded and executed by the processor to implement the control method for the self - moving device provided in the first aspect.

[0038] In a fourth aspect, a computer - readable storage medium is provided. A program is stored in the storage medium, and the program is loaded and executed by the processor to implement the control method for the self - moving device provided in the first aspect.

[0039] In a fifth aspect, a self - moving device is provided, including:

[0040] A moving component for driving the self - moving device to move;

[0041] A moving drive component for driving the movement of the moving component;

[0042] An image acquisition component installed on the self - moving device for acquiring environmental images in the traveling direction;

[0043] A control component communicatively connected to the moving drive component and the image acquisition component, and the control component is communicatively connected to the memory; a program is stored in the memory, and the program is loaded and executed by the control component to implement the control method for the self - moving device provided in the first aspect.

[0044] The beneficial effects of the present application are as follows: by obtaining the edge information of the working area where the self-mobile device is located; obtaining the environmental images collected during movement; obtaining the access door information based on the environmental images; dividing the independent areas within the working area based on the access door information and the edge information; it can solve the problem that the prior art cannot divide the working area; it can divide the working area and achieve the effect of personalized work according to the divided independent areas. In addition, since the access door can be an open virtual door, therefore, the information of the virtual door and the edge information can be combined to obtain the area boundary of each independent area, thereby dividing the corresponding independent area, and the area division of the open door scenario can be realized, and the accuracy of area division can be improved.

[0045] The above description is only an overview of the technical solution of the present application. In order to be able to understand the technical means of the present application more clearly and implement it in accordance with the content of the specification, the following takes the preferred embodiments of the present application and combines the accompanying drawings to describe in detail as follows. Brief Description of the Drawings

[0046] Figure 1 is a schematic structural diagram of a self-mobile device provided by an embodiment of the present application;

[0047] Figure 2 is a flowchart of a control method of a self-mobile device provided by an embodiment of the present application;

[0048] Figure 3 is a flowchart of a control method of a self-mobile device provided by another embodiment of the present application;

[0049] Figure 4 is a schematic diagram of the control of a self-mobile device provided by an embodiment of the present application;

[0050] Figure 5 is a block diagram of a control device of a self-mobile device provided by an embodiment of the present application;

[0051] Figure 6 is a block diagram of a control device of a self-mobile device provided by an embodiment of the present application. Detailed Description of the Embodiments

[0052] The following combines the accompanying drawings and embodiments to further describe in detail the specific implementation manners of the present application. The following embodiments are used to illustrate the present application, but are not used to limit the scope of the present application.

[0053] First, several terms related to the present application are introduced below.

[0054] Model compression: refers to a way of reducing the parameter redundancy in the trained network model, thereby reducing the storage occupancy, communication bandwidth, and computational complexity of the network model.

[0055] Model compression includes, but is not limited to: model pruning, model quantization, and / or low-rank factorization.

[0056] Model pruning: It refers to the search process for the optimal network structure. The model pruning process includes: 1. Training the network model; 2. Pruning unimportant weights or channels; 3. Fine-tuning or retraining the pruned network. Among them, the second step usually uses iterative layer-by-layer pruning, fast fine-tuning, or weight reconstruction to maintain accuracy.

[0057] Quantization: Quantizing a model is a general term for a model acceleration method. It is the process of representing floating-point data within a limited range (such as 32 bits) using a data type with fewer bits, thereby achieving goals such as reducing the model size, reducing the model memory consumption, and accelerating the model inference speed.

[0058] Low-rank factorization: Decompose the weight matrix of the network model into multiple small matrices. The computational amount of the small matrices is smaller than that of the original matrix, so as to reduce the computational amount of the model and the memory occupied by the model.

[0059] YOLO model: One of the basic network models, it is a neural network model that can achieve object localization and recognition through a Convolutional Neural Networks (CNN) network. The YOLO model includes YOLO, YOLO v2, and YOLO v3. Among them, YOLO v3 is another object detection algorithm in the YOLO series after YOLO and YOLO v2, and it is an improvement based on YOLO v2. And YOLO v3-tiny is a simplified version of YOLO v3. By removing some feature layers on the basis of YOLO v3, it achieves the effect of reducing the computational amount of the model and faster operation.

[0060] MobileNet model: A network model whose basic unit is depthwise separable convolution. Among them, depthwise separable convolution can be decomposed into depthwise convolution (DW) and pointwise convolution (PW). DW is different from standard convolution. For standard convolution, its convolution kernel is used on all input channels, while DW uses different convolution kernels for each input channel, that is, one convolution kernel corresponds to one input channel. And PW is ordinary convolution, but it uses a 1x1 convolution kernel. For depthwise separable convolution, it first uses DW to perform convolution on different input channels respectively, and then uses PW to combine the above outputs. In fact, the overall calculation result is approximately the same as that of a standard convolution process, but it will greatly reduce the computational amount and the number of model parameters.

[0061] Probability Model: It refers to a mathematical model of the joint probability distribution or conditional probability distribution of the observed variable pair (X, Y) under the system parameter w, that is, P(Y, X|w). Among them, the system parameter w can be determined through a large number of observed data (or called training data). Probability models include but are not limited to: Bayesian probability models (Bayesian Models), Markov probability models (Markov Models), etc., which are not listed one by one in this application.

[0062] Figure 1 It is a schematic structural diagram of a self-mobile device provided by an embodiment of the present application, as Figure 1 shown. The system at least includes: a control component 110 and an image acquisition component 120 communicatively connected to the control component 110.

[0063] The image acquisition component 120 is used to acquire the environmental image 130 during the movement of the self-mobile device; and send the environmental image 130 to the control component 110. Optionally, the image acquisition component 120 can be implemented as a camera, a video camera, etc., and the implementation manner of the image acquisition component 120 is not limited in this embodiment.

[0064] Optionally, the field of view angle of the image acquisition component 120 is 120° in the horizontal direction and 60° in the vertical direction; of course, the field of view angle can also be other values, and the value of the field of view angle of the image acquisition component 120 is not limited in this embodiment. The field of view angle of the image acquisition component 120 can ensure that the environmental image 130 in the traveling direction of the self-mobile device can be acquired.

[0065] In addition, the number of the image acquisition components 120 can be one or more, and the number of the image acquisition components 120 is not limited in this embodiment.

[0066] The control component 110 is used to control the self-mobile device. For example: controlling the start and stop of the self-mobile device; controlling the start and stop of each component (such as the image acquisition component 120) in the self-mobile device, etc.

[0067] In this embodiment, the control component 110 is communicatively connected to the memory; a program is stored in the memory, and the program is loaded and executed by the control component 110 to at least implement the following steps: used to obtain the edge information 140 of the working area where the self-mobile device is located; obtain the environmental image 130 collected during movement; obtain the access door information based on the environmental image 130; divide the independent area 150 in the working area based on the access door information and the edge information 140. In other words, the program is loaded and executed by the control component 110 to implement the control method of the self-mobile device provided by the present application.

[0068] The independent area 150 refers to an area in the working area with different attributes from other areas. For example: the bedroom area, living room area, dining room area, kitchen area, etc. in a room. Since different independent areas are usually divided by passage doors and walls, among which, the boundary line between the wall and the ground can be obtained through edge information, and the position of the passage door can be obtained through passage door information, therefore, area division can be carried out by combining edge information and passage door information.

[0069] Among them, the edge information 140 refers to the information of the ground boundary of each independent area, and this edge information includes the position and length of the corresponding ground boundary. The passage door information refers to the information of the passage doors in each independent area, and the passage door information includes the position information of the door frame and / or the door in the working area. A passage door refers to a door for people, self-moving devices and / or other objects to enter and exit a certain independent area. The passage door can be an open virtual door (i.e., a door without isolation objects such as door panels), or a physical door, and the type of the passage door is not limited in this embodiment.

[0070] It should be added that the position information of the door frame and / or the door in the working area means: in the direction of the vertical projection onto the ground, the projection position of the door frame and / or the door in the geographical location in the working area.

[0071] Optionally, after dividing the independent area 150 in the working area, the control component 110 can also implement the following steps: determining the scene prediction result 160 corresponding to the independent area 150 based on the environmental image 130; determining the scene type 170 of the independent area according to the scene prediction result 160 of the independent area 150.

[0072] The scene type 170 is used to indicate the type of the independent area where the self-moving device is currently located, and the division method of the scene type 170 is set according to the working area of the self-moving device. For example: if the working area of the self-moving device is a room, then the scene type 170 includes: bedroom type, kitchen type, study type, bathroom type, etc., and the division method of the scene type 170 is not limited in this embodiment.

[0073] It should be added that in this embodiment, the self-moving device may also include other components, such as: a moving component (such as: wheels) for driving the self-moving device to move, a moving drive component (such as: a motor) for driving the moving component to move, etc. Among them, the moving drive component is communicatively connected to the control component 110, and under the control of the control component 110, the moving drive component operates and drives the moving component to move, so as to realize the overall movement of the self-moving device. The components included in the self-moving device are not listed one by one in this embodiment.

[0074] In addition, the self - moving device can be a floor - cleaning robot, an automatic lawn mower, or other devices with an automatic driving function. The present application does not limit the type of the self - moving device.

[0075] In this embodiment, by combining the edge information and the passage door information to divide the working area into multiple independent areas, the problem that the prior art cannot divide the working area can be solved; the working area can be divided, and the effect of personalized work according to the divided independent areas can be achieved. In addition, since the passage door can be an open virtual door, therefore, the information of the virtual door and the edge information can be combined to obtain the area boundary of each independent area, so as to divide the corresponding independent area, and the area division of the open - door scenario can be realized, improving the accuracy of area division.

[0076] The object recognition method of the self - moving device provided by the present application will be introduced in detail below.

[0077] Figure 2 is the flowchart of the object recognition method of the self - moving device. Figure 2 In, the object recognition method of the self - moving device is used for Figure 1 the self - moving device shown in, and taking the control component 110 as the execution subject of each step as an example for illustration, the method at least includes the following steps:

[0078] Step 201, obtain the edge information of the working area where the self - moving device is located.

[0079] The edge information refers to the information of the ground boundary of each independent area, and this edge information includes the position and length of the corresponding ground boundary. The edge information is the path information obtained when the self - moving device travels along the edge during the moving process.

[0080] Optionally, the self - moving device has the function of traveling along the edge. In this function, if the self - moving device travels along the boundary formed by the wall and the ground, the edge information of this boundary can be obtained; if the self - moving device travels along the boundary formed by an object (such as a cabinet, a table, a bed, etc.) and the ground, the edge information of this boundary can be obtained.

[0081] Step 202, obtain the environmental image collected during the movement.

[0082] The self - moving device records the working time (or time stamp) during the movement in the working area and the environmental image corresponding to this working time. After the work in the working area is completed, the environmental image is read.

[0083] In an example, after the self - moving device completes the current work, according to the start time and end time of the current work, the environmental image corresponding to the time period from the start time to the end time is read from the stored environmental images.

[0084] Step 203: Obtain access door information based on the environmental image.

[0085] Identify whether the environmental image includes an image of the access door; when the environmental image includes an image of the access door, obtain the position information of the access door in the working area.

[0086] Herein, the access door refers to a passage for the self - moving device to enter or leave the independent area. The access door can be a door frame, a fence opening, etc. The type of the access door is not limited in this embodiment. In one example, the access door includes the door frame in the working area.

[0087] In one example, an image recognition model is stored in the self - moving device; the self - moving device inputs the environmental image into the image recognition model to obtain an object recognition result, and the object recognition result includes the attribute information of the target object. When the object recognition result includes the access door information, it is determined that the environmental image includes an image of the access door; when the object recognition result does not include the access door information, it is determined that the environmental image does not include an image of the access door. Herein, the target object includes the access door.

[0088] The image recognition model is trained based on a small network model using sample environmental images and sample object results.

[0089] Optionally, in order to reduce the hardware requirements of the self - moving device for the image recognition process, the computing resources occupied when the image recognition model runs are lower than the maximum computing resources provided by the self - moving device. The image recognition model is trained based on a small network model using training data. Herein, the training data includes the training images of each object in the working area of the self - moving device and the recognition results of each training image. Herein, the small network model refers to a network model with the number of model layers less than the first value; and / or the number of nodes in each layer less than the second value. Both the first value and the second value are small integers. For example: the small network model is a miniature YOLO model; or a MobileNet model. Of course, the small network model can also be other models, which are not listed one by one in this embodiment.

[0090] Optionally, in order to further compress the computing resources occupied when the image recognition model runs, after the image recognition model is trained, the self - moving device can also perform model compression processing on the image recognition model. The model compression processing includes, but is not limited to: model pruning, model quantization, and / or low - rank decomposition, etc.

[0091] Herein, the object recognition result includes the attribute information of the target object, such as: the type of the object, the size of the object, and the position information of the object in the environmental image. In addition to the access door, the target object can also include household items such as beds, tables, sofas, etc. The type of the target object is not limited in this embodiment.

[0092] Optionally, obtain the position information of the access door in the working area, including: obtaining the first distance between the access door and the self-moving device in the environmental image output by the image recognition model; obtaining the second distance between the edge information indicating the boundary when the environmental image is collected; determining the position of the access door relative to the boundary indicated by the edge information according to the first distance and the second distance to obtain the position information of the access door. Alternatively, a positioning component is installed on the self-moving device, and the self-moving device obtains the first distance between the access door and the self-moving device in the environmental image output by the image recognition model; obtains the positioning information obtained by the positioning component when the environmental image is collected; and obtains the position information of the access door according to the first distance and the positioning information. Of course, the method for the self-moving device to obtain the position information of the access door can also be other methods, which are not listed one by one in this embodiment.

[0093] Step 204, divide the independent areas in the working area based on the access door information and the edge information.

[0094] Since one or more independent areas in the working area may not be completely enclosed but are connected to other areas in the working area, such as areas like an open kitchen, etc., at this time, the independent area cannot be distinguished from other areas only based on the edge information of the self-moving device. Based on this, in this embodiment, by combining the edge information with the access door information, an open independent area can be determined, improving the accuracy of area division.

[0095] Dividing the independent areas in the working area based on the access door information and the edge information includes: obtaining the position information of the corresponding access door indicated by the access door information in the working area; combining the edge information and the position information to obtain the combined boundary information; and dividing each closed area formed by the combined boundary information into corresponding independent areas.

[0096] Among them, the edge information is obtained by the self-moving device driving along the edge.

[0097] For example: the self-moving device is a sweeping robot. After the sweeping robot finishes cleaning the entire house of the user, it obtains the edge information of the house; then, it obtains the environmental images collected during the cleaning process and recognizes each environmental image; when the environmental image includes an image of the access door, it obtains the position information of the access door; and combining the position information with the edge information can obtain multiple closed figures, and each closed figure corresponds to an independent area.

[0098] In summary, the control method for the self - moving device provided in this embodiment obtains the edge information of the working area where the self - moving device is located; obtains the environmental images collected during movement; obtains the access door information based on the environmental images; divides the independent areas within the working area based on the access door information and the edge information, which can solve the problem that the prior art cannot divide the working area; can divide the working area and achieve the effect of personalized work according to the divided independent areas. In addition, since the access door can be an open virtual door, the information of the virtual door and the edge information can be combined to obtain the area boundary of each independent area, thereby dividing the corresponding independent areas, and the area division of the open - door scenario can be realized, improving the accuracy of area division.

[0099] In addition, by compressing the image recognition model, an image recognition model for identifying access doors is obtained, which can further reduce the computing resources occupied during the operation of the image recognition model, improve the recognition speed, and reduce the hardware requirements for the self - moving device.

[0100] Optionally, after obtaining multiple independent areas within the working area, the self - moving device can also identify the scene type of each independent area. At this time, after step 204, referring to Figure 3 , the control method of the self - moving device further includes the following steps:

[0101] Step 301, determine the scene prediction result corresponding to the independent area based on the environmental image.

[0102] The scene prediction result is used to indicate the scene type predicted by the self - moving device based on the relevant information of a single independent area. The scene prediction result can be one or more scene types.

[0103] The scene type is used to indicate the type of the independent area where the self - moving device is currently located. The division method of the scene type is set according to the working area of the self - moving device. For example, if the working area of the self - moving device is a room, the scene types include: bedroom type, kitchen type, study type, bathroom type, etc. The division method of the scene type is not limited in this embodiment.

[0104] In one example, determining the scene prediction result corresponding to the independent area based on the environmental image includes: obtaining the image recognition model; for each independent area, input the environmental image corresponding to the independent area into the image recognition model to obtain the object recognition result; obtain the scene recognition model; input the object recognition result into the scene recognition model to obtain the scene prediction result, and the scene prediction result includes at least one predicted scene type of the independent area.

[0105] The relevant description of the image recognition model can be found in step 203, and will not be elaborated here in this embodiment.

[0106] The scene recognition model is trained based on a probability model using the sample attribute information of the object and the sample scene.

[0107] Optionally, the scene prediction result output by the scene recognition model includes multiple scene types. At this time, the scene recognition model also outputs the confidence level corresponding to each scene type. The confidence level is used to indicate the accuracy of each output scene type.

[0108] Step 302, determine the scene type of the independent area according to the scene prediction result of the independent area.

[0109] In one example, determining the scene type of the independent area according to the scene prediction result of the independent area includes: obtaining the pose information of each independent area; combining the scene area result of each independent area and the pose information of each independent area, and determining the scene type of each independent area according to the preset probability distribution strategy. Among them, the probability distribution strategy is used to determine, for each target scene type, the independent area with the highest probability of the scene type being the target scene type from each independent area.

[0110] The pose information includes the position information and direction information of the corresponding independent area in the working area. Among them, the direction information can be the direction of the access door in the corresponding independent area.

[0111] When determining the scene type of the independent area, the self - moving device can determine the scene type of each independent area according to the scene prediction results of multiple independent areas according to the preset probability distribution strategy;

[0112] Schematically, the probability distribution strategy is: there is corresponding template pose information for each scene type; compare the pose information of the independent area with the template pose information of each scene type to obtain the pose comparison result; for each target scene type, multiply the scene prediction result corresponding to the target scene type and the pose comparison result by the corresponding weight and sum them to obtain the probability result; determine the type of the independent area with the highest probability result as the scene type.

[0113] To more clearly understand the control method of the self - moving device provided in this application, an example of this method is given below for illustration. Refer to Figure 4 , after the self - moving device finishes working in the working area, it obtains the edge information in the working area; inputs the environmental image collected by the image acquisition component during the working process into the image recognition model 41 to obtain the object information 42; combines the access door information and the edge information in the object information 42 to divide the working area to obtain multiple independent areas 150; inputs the object information into the scene recognition model to obtain the scene prediction result 160 of each independent area; combines the scene prediction results 160 of multiple independent areas and obtains the scene type 170 of each independent area based on the probability distribution strategy.

[0114] In summary, for the control method of the self - moving device provided in this embodiment, by determining the scene prediction result of the corresponding independent area based on the environmental image, and determining the scene type of the independent area according to the scene prediction result of the independent area, the self - moving device can identify the scene type of each independent area in the entire working area, provide more information for the user, and make the self - moving device more intelligent.

[0115] Figure 5 is a block diagram of a control device for a self - moving device provided by an embodiment of the present application. In this embodiment, it is described by taking the device as an example of the control component 110 in the control system of the self - moving device shown in Figure 1 The device at least includes the following several modules: a first information acquisition module 510, an environmental image acquisition module 520, a second information acquisition module 530, and an area division control module 540.

[0116] The first information acquisition module 510 is configured to acquire the edge information of the working area where the self - moving device is located;

[0117] The environmental image acquisition module 520 is configured to acquire the environmental image collected during movement;

[0118] The second information acquisition module 530 is configured to obtain the access door information based on the environmental image;

[0119] The area division control module 540 is configured to divide the independent areas in the working area based on the access door information and the edge information.

[0120] For related details, refer to the above - mentioned method embodiment.

[0121] It should be noted that: when the control device of the self - moving device provided in the above - mentioned embodiment controls the self - moving device, only the division of the above - mentioned each functional module is used for illustration. In practical applications, the above functions can be allocated to different functional modules according to needs, that is, the internal structure of the control device of the self - moving device is divided into different functional modules to complete all or part of the functions described above. In addition, the control device of the self - moving device provided in the above - mentioned embodiment and the method embodiment of the control method of the self - moving device belong to the same concept. For the specific implementation process, refer to the method embodiment, and details are not described here again.

[0122] Figure 6 is a block diagram of a control device for a self - moving device provided by an embodiment of the present application. The device can be Figure 1 the self - moving device shown in. The device at least includes a processor 601 and a memory 602.

[0123] The processor 601 may include one or more processing cores, such as a 4-core processor, an 8-core processor, etc. The processor 601 may be implemented in at least one hardware form of DSP (Digital Signal Processing), FPGA (Field-Programmable Gate Array), or PLA (Programmable Logic Array). The processor 601 may also include a main processor and a coprocessor. The main processor is a processor used to process data in the wake state, also known as the CPU (Central Processing Unit); the coprocessor is a low-power processor used to process data in the standby state. In some embodiments, the processor 601 may further include an AI (Artificial Intelligence) processor, which is used to process computational operations related to machine learning.

[0124] The memory 602 may include one or more computer-readable storage media, and the computer-readable storage media may be non-transitory. The memory 602 may further include high-speed random access memory and non-volatile memory, such as one or more disk storage devices and flash storage devices. In some embodiments, the non-transitory computer-readable storage media in the memory 602 is used to store at least one instruction, and the at least one instruction is used to be executed by the processor 601 to implement the control method of the self-mobile device provided in the method embodiments of the present application.

[0125] In some embodiments, the control device of the self-mobile device may optionally further include: a peripheral device interface and at least one peripheral device. The processor 601, the memory 602, and the peripheral device interface may be connected through a bus or signal lines. Each peripheral device may be connected to the peripheral device interface through a bus, signal lines, or a circuit board. Schematically, the peripheral devices include, but are not limited to: a radio frequency circuit, a touch display screen, an audio circuit, and a power supply, etc.

[0126] Of course, the control device of the self-mobile device may also include fewer or more components, and this embodiment does not limit this.

[0127] Optionally, the present application further provides a computer-readable storage medium, and a program is stored in the computer-readable storage medium. The program is loaded and executed by the processor to implement the control method of the self-mobile device in the above method embodiments.

[0128] Optionally, the present application also provides a computer product, which includes a computer-readable storage medium. A program is stored in the computer-readable storage medium and is loaded and executed by a processor to implement the control method of the self-mobile device in the above method embodiments.

[0129] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as the scope recorded in this specification.

[0130] The above embodiments only represent several implementation manners of the present application. The description is relatively specific and detailed, but it should not be understood as a limitation on the scope of the invention patent. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present application, several modifications and improvements can still be made, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application should be subject to the appended claims.

Claims

1. A control method for a self - moving device, characterized in that, the method includes: Obtaining edge information of the working area where the self - moving device is located; Obtaining an environmental image collected during movement; Based on the environmental image, obtaining the position information of the access door in the working area, including: obtaining the first distance between the access door in the environmental image output by the image recognition model and the self - moving device; obtaining the second distance between the self - moving device when collecting the environmental image and the boundary indicated by the edge information; determining the position of the access door relative to the boundary indicated by the edge information according to the first distance and the second distance to obtain the position information of the access door; Or, a positioning component is installed on the self - moving device, and the self - moving device obtains the first distance between the access door in the environmental image output by the image recognition model and the self - moving device; obtaining the positioning information obtained by the positioning component when collecting the environmental image; obtaining the position information of the access door according to the first distance and the positioning information; Based on the access door information and the edge information, dividing independent areas within the working area.

2. The method according to claim 1, characterized in that, after dividing the independent areas within the working area based on the access door information and the edge information, it further includes: Determining a scene prediction result corresponding to the independent area based on the environmental image; Determining the scene type of the independent area according to the scene prediction result of the independent area.

3. The method according to claim 1, characterized in that, dividing the independent areas within the working area based on the access door information and the edge information includes: Obtaining the position information of the corresponding access door indicated by the access door information in the working area; Combining the edge information and the position information to obtain combined boundary information; Dividing each closed area formed by the combined boundary information into corresponding independent areas.

4. The method according to claim 1, characterized in that, obtaining the access door information based on the environmental image includes: Identifying whether the environmental image includes an image of an access door; When the environmental image includes an image of an access door, obtaining the position information of the access door in the working area.

5. The method according to claim 4, characterized in that, the access door includes a door frame in the working area.

6. The method according to claim 2, characterized in that, determining the scene type of the independent area according to the scene prediction result of the independent area includes: Obtaining the pose information of each independent area, where the pose information includes the position information and direction information of the corresponding independent area within the working area; Combining the scene area result of each independent area and the pose information of each independent area, and determining the scene type of each independent area according to a pre - set probability distribution strategy; wherein, the probability distribution strategy is used to determine, for each target scene type, the independent area with the highest probability of having the scene type as the target scene type from each independent area.

7. The method according to claim 2, characterized in that, Determining a scene prediction result corresponding to an independent region based on the environmental image includes: Obtain an image recognition model, where the computing resources occupied during the operation of the image recognition model are lower than the maximum computing resources provided by the self - moving device; For each independent region, input the environmental image corresponding to the independent region into the image recognition model to obtain an object recognition result, where the object recognition result includes attribute information of the target object; Obtain a scene recognition model, where the scene recognition model is trained using sample attribute information of objects and sample scene types; Input the object recognition result into the scene recognition model to obtain a scene prediction result, where the scene prediction result includes at least one predicted scene type of the independent region.

8. The method according to claim 6, wherein, The probability distribution strategy includes: For each scene type, there is corresponding template pose information. Compare the pose information of the independent region with the template pose information of each scene type to obtain a pose comparison result; For each target scene type, multiply the scene prediction result and the pose comparison result corresponding to the target scene type by the corresponding weights and sum them to obtain a probability result; Determine the type of the independent region with the highest probability result as the scene type.

9. A control device for a self - moving device, wherein, The device includes a processor and a memory; a program is stored in the memory, and the program is loaded and executed by the processor to implement the control method of the self - moving device according to any one of claims 1 to 8.

10. A computer - readable storage medium, wherein, A program is stored in the storage medium, and when the program is executed by a processor, it is used to implement the control method of the self - moving device according to any one of claims 1 to 8.

11. A self - moving device, wherein, including: A moving component for driving the self - moving device to move; A moving driving component for driving the moving component to move; An image acquisition component installed on the self - moving device for acquiring environmental images in the traveling direction; A control component communicatively connected to the moving driving component and the image acquisition component, and the control component is communicatively connected to the memory; a program is stored in the memory, and the program is loaded and executed by the control component to implement the control method of the self - moving device according to any one of claims 1 to 8.

Images