Clustering method and device for three-dimensional map of cleaning robot, equipment and medium
By clustering the three-dimensional map of the cleaning robot, the results of different clustering types are displayed, and the problem of single display of three-dimensional maps in the existing technology is solved, and user experience and environmental understanding are improved.
Patent Information
- Application Number
- CN202411147496.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-20
- Publication Date
- 2024-11-22
AI Technical Summary
The three-dimensional map display method of existing cleaning robots is single, and cannot meet the diverse viewing needs of users.
By obtaining the three-dimensional map of the cleaning robot and using the clustering algorithm to cluster point clouds or geometric primitives, marking and displaying them according to different cluster types (surface clustering, regional clustering, semantic clustering, etc.), the three-dimensionality is improved Diversity of maps.
It realizes the diversity display of three-dimensional maps, meets users' diverse viewing needs, and improves users' intuitive understanding of the working environment of cleaning robots.
Smart Images

Figure CN119006866A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of robots, and in particular to a clustering method, device, equipment and medium for a three-dimensional map of a cleaning robot. Background Art
[0002] With the development of computer technology and artificial intelligence technology, various cleaning robots with intelligent systems have emerged, such as sweeping machines, mopping robots, vacuum cleaners, lawn mowers, etc. These robots can automatically move in a certain area and perform operations such as cleaning or clearing without user operation. At present, the three-dimensional map display method of the cleaning robot's working environment is single and cannot meet the viewing needs of users. Summary of the invention
[0003] A series of simplified concepts are introduced in the Summary of the Invention, which will be further described in detail in the Detailed Description of the Invention. The Summary of the Invention does not mean to attempt to define the key features and essential technical features of the claimed technical solution, nor does it mean to attempt to determine the scope of protection of the claimed technical solution.
[0004] In a first aspect, an embodiment of the present invention provides a clustering method for a three-dimensional map of a cleaning robot, comprising:
[0005] Obtaining a three-dimensional map of the cleaning robot and at least one clustering type;
[0006] Based on the clustering type, clustering the point cloud or geometric primitives of the three-dimensional map using a clustering algorithm to obtain a clustered three-dimensional map;
[0007] The clustered three-dimensional map is displayed.
[0008] Optionally, the at least one clustering type includes a face clustering type, and based on the clustering type, clustering the point cloud or geometric primitives of the three-dimensional map using a clustering algorithm to obtain a clustered three-dimensional map includes:
[0009] Based on the face clustering type, clustering the point cloud or geometric primitives in the three-dimensional map using a clustering algorithm to obtain planes and curved surfaces in the three-dimensional map;
[0010] Based on the smoothness of the planes and the curved surfaces, the planes or curved surfaces are clustered using the clustering algorithm to determine planes and / or curved surfaces belonging to the same category; the planes and / or curved surfaces of the same category are marked in the three-dimensional map according to a preset marking method to obtain a clustered three-dimensional map.
[0011] Optionally, the at least one clustering type includes a regional clustering type, and based on the clustering type, clustering the point cloud or geometric primitives of the three-dimensional map using a clustering algorithm to obtain a clustered three-dimensional map includes:
[0012] Based on the region clustering type, clustering the point cloud or geometric primitives in the three-dimensional map using a clustering algorithm to obtain at least one target region on the three-dimensional map, wherein the target region is a region of interest to the user;
[0013] Each target area in the three-dimensional map is marked according to a preset marking method to obtain a clustered three-dimensional map.
[0014] Optionally, the at least one clustering type includes a semantic clustering type, and based on the clustering type, clustering the point cloud or geometric primitives of the three-dimensional map using a clustering algorithm to obtain a clustered three-dimensional map includes:
[0015] Based on the semantic clustering type, clustering the point cloud or geometric primitives in the three-dimensional map using a clustering algorithm to obtain semantic information of the three-dimensional map;
[0016] In the three-dimensional map, each piece of semantic information is marked according to a preset marking method to obtain a clustered three-dimensional map.
[0017] Optionally, the clustering algorithm includes a K-means algorithm, a GMM algorithm, a DBSCAN algorithm, a spectral clustering algorithm or a semantic clustering algorithm of deep learning.
[0018] Optionally, the preset marking method includes color marking.
[0019] Optionally, after displaying the clustered three-dimensional map, the step further includes:
[0020] The clustering type and the corresponding clustered three-dimensional map are stored.
[0021] In a second aspect, an embodiment of the present invention provides a clustering device for a three-dimensional map of a cleaning robot, comprising:
[0022] An acquisition module, used to acquire a three-dimensional map of the cleaning robot and at least one clustering type;
[0023] A clustering module, used for clustering the point cloud or geometric primitives of the three-dimensional map using a clustering algorithm based on the clustering type to obtain a clustered three-dimensional map;
[0024] The display module is used to display the clustered three-dimensional map.
[0025] In a third aspect, an embodiment of the present invention provides an electronic device, including a processor and a memory, wherein the memory is used to store at least one executable instruction, and the executable instruction enables the processor to execute the steps of the above-mentioned method for clustering the three-dimensional map of the cleaning robot.
[0026] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium storing computer program instructions, which, when called and executed by a processor, implement the steps of the above-mentioned method for clustering the three-dimensional map of the cleaning robot.
[0027] According to an embodiment of the present invention, a clustering method, device, equipment and medium for a three-dimensional map of a cleaning robot are provided. The method obtains a three-dimensional map of the cleaning robot and at least one clustering type; based on the clustering type, a clustering algorithm is used to cluster the point cloud or geometric primitives of the three-dimensional map to obtain a clustered three-dimensional map; and the clustered three-dimensional map is displayed. Thus, the three-dimensional map can be clustered according to different clustering types and then displayed, thereby improving the diversity of the three-dimensional map display and meeting the viewing needs of users. BRIEF DESCRIPTION OF THE DRAWINGS
[0028] The following drawings of the present invention are used as part of the embodiments of the present invention for understanding the present invention. The embodiments of the present invention are shown in the drawings and the description thereof is used to explain the principle of the present invention.
[0029] In the attached figure:
[0030] Figure 1 A schematic diagram of a system structure according to an optional embodiment of the present invention;
[0031] Figure 2 is a three-dimensional schematic diagram of a cleaning robot according to an optional embodiment of the present invention;
[0032] Figure 3 A bottom view of a cleaning robot according to an optional embodiment of the present invention;
[0033] Figure 4 is a perspective view of a wet cleaning system according to an optional embodiment of the present invention;
[0034] Figure 5 A flowchart of a clustering method for a three-dimensional map of a cleaning robot according to an optional embodiment of the present invention;
[0035] Figure 6 is a flowchart of step S502 according to an optional embodiment of the present invention;
[0036] Figure 7 is a flowchart of step S502 according to another optional embodiment of the present invention;
[0037] Figure 8 is a flowchart of step S502 according to another optional embodiment of the present invention;
[0038] Fig. 9 It is a structural diagram of a clustering device for a three-dimensional map of a cleaning robot according to an optional embodiment of the present invention. DETAILED DESCRIPTION
[0039] In the following description, a large number of specific details are provided to provide a more thorough understanding of the present invention. However, it is apparent to those skilled in the art that the present invention can be implemented without one or more of these details. In other examples, in order to avoid confusion with the present invention, some technical features well known in the art are not described.
[0040] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit exemplary embodiments according to the present invention. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should also be understood that when the terms "comprise" and / or "include" are used in this specification, it indicates the presence of the features, wholes, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components and / or combinations thereof.
[0041] Now, exemplary embodiments according to the present invention will be described in more detail with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in a variety of different forms and should not be interpreted as being limited to the embodiments set forth herein. It should be understood that these embodiments are provided to make the disclosure of the present invention thorough and complete, and to fully convey the concept of these exemplary embodiments to those of ordinary skill in the art.
[0042] Figure 1 An exemplary system architecture of the three-dimensional map construction method of the cleaning robot disclosed in the present invention is shown. Figure 1 As shown, the system architecture may include terminal devices, networks, servers, and cleaning robots. The terminal device may be a variety of electronic devices with a display screen and supporting input and output, including but not limited to smartphones, tablet computers, laptops, desktop computers, AR headsets, and the like. The network is a medium used to provide a communication link between the terminal device and the server. The network may include various connection types, such as wired, wireless communication links, or fiber optic cables, and the like. The server may be a server or server cluster that provides various services, such as a background processing server that provides support for map modeling of data sent by a sweeping robot to perceive the working environment.
[0043] The user can use the terminal device to interact with the server through the network to receive or send data, etc. For example, the server sends the constructed three-dimensional map to the mobile phone through the network, and the user can view the three-dimensional map through the mobile phone.
[0044] It should be understood that Figure 1 The number of terminal devices 2, networks 4 and servers 3 in the embodiment is only illustrative. According to the implementation requirements, there may be any number of terminal devices 2, networks 4 and servers 3. The cleaning robot may be a sweeping robot, a mopping robot, a floor polishing robot or a weeding robot. For ease of description, the present embodiment takes the sweeping robot 10 as an example to describe the technical solution of the present disclosure.
[0045] like Figure 2 and Figure 3 As shown, the cleaning robot 10 may include a machine body 110, a sensing module 120, a controller, a driving module, a cleaning system 150, an energy system and a human-machine interaction module 130. Figure 1 As shown, the machine body 110 includes a forward portion 111 and a rearward portion 112, and has an approximately circular shape (both the front and the rear are circular), and may also have other shapes, including but not limited to an approximately D-shape with a front and rear circle and a rectangular or square shape with a front and rear.
[0046] like Figure 2 As shown, the perception module 120 includes a position determination device 121 located on the machine body 110, a collision sensor disposed on a front collision structure 122 of the forward portion 111 of the machine body 110, a short-distance sensor (wall sensor) located on the side of the machine, a carpet detection device disposed at the lower portion of the machine body 110, and a magnetometer, accelerometer, gyroscope, odometer and other sensing devices disposed inside the machine body 110, which are used to provide the controller with various position information and motion state information of the machine. The position determination device 121 includes but is not limited to a camera and a laser distance measuring device (LDS, full name Laser Distance Sensor). In some preferred implementations, the position determination device 121 (such as a camera, a laser sensor) is located on the front side of the body 110, that is, at the very front end of the forward portion 111, so as to be able to more accurately sense the environment in front of the cleaning robot and achieve precise positioning.
[0047] like Figure 2As shown, the forward portion 111 of the machine body 110 can carry a front collision structure 122. During the cleaning process, when the driving wheel module 141 propels the cleaning robot 10 to walk on the ground, the front collision structure 122 detects one or more events in the driving path of the cleaning robot 10 via a sensor system arranged thereon, such as a collision sensor or a proximity sensor (infrared sensor). The cleaning robot 10 can control the driving module to enable the cleaning robot 10 to respond to the event, such as performing an obstacle avoidance operation away from the obstacle, through the event detected by the front collision structure 122, such as an obstacle or a wall.
[0048] The controller is arranged on the circuit board in the machine body 110, and includes a computing processor, such as a central processing unit, an application processor, which communicates with a non-temporary memory, such as a hard disk, a flash memory, and a random access memory. The application processor uses a positioning algorithm, such as SLAM (Simultaneous Localization And Mapping), to draw a real-time map of the environment where the cleaning robot 10 is located based on the obstacle information fed back by the laser ranging device. In addition, the distance information and speed information fed back by the sensors, carpet detection devices, magnetometers, accelerometers, gyroscopes, odometers and other sensor devices arranged on the front collision structure 122 are combined to comprehensively judge the current working state and position of the cleaning robot 10, as well as the current posture of the cleaning robot 10, such as passing the threshold, getting on the carpet, the dust box is full, being picked up, etc., and specific next action strategies are given for different situations, so that the cleaning robot 10 has better cleaning performance and user experience.
[0049] like Figure 3As shown, the driving module can manipulate the machine body 110 to travel across the ground based on a driving command with distance and angle information. The driving module includes a main driving wheel module, which can control the left wheel 140 and the right wheel 141. In order to more accurately control the movement of the machine, it is preferred that the main driving wheel module includes a left driving wheel module and a right driving wheel module respectively. The left and right driving wheel modules are arranged along the transverse axis defined by the machine body 110. In order for the cleaning robot 10 to be able to move more stably or have stronger movement ability on the ground, the cleaning robot 10 may include one or more driven wheels 142, and the driven wheels 142 include but are not limited to universal wheels. The main driving wheel module includes a driving motor and a control circuit for controlling the driving motor, and the main driving wheel module can also be connected to a circuit and an odometer for measuring the driving current. And the left wheel 140 and the right wheel 141 may have a biased drop suspension system, which is fastened in a movable manner, for example, attached to the machine body 110 in a rotatable manner, and receives a spring bias biased downward and away from the machine body 110. The spring bias allows the drive wheels to maintain contact and traction with the ground with a certain ground force, while the cleaning elements of the cleaning robot 10 also contact the ground with a certain pressure.
[0050] The energy system includes rechargeable batteries, such as nickel-metal hydride batteries and lithium batteries. The rechargeable batteries can be connected to a charging control circuit, a battery pack charging temperature detection circuit, and a battery undervoltage monitoring circuit, which are then connected to a single-chip microcomputer control circuit. The host is charged by connecting to a charging pile through a charging electrode 160 disposed on the side or below the fuselage.
[0051] The human-computer interaction module 130 includes buttons on the host panel, which are used by the user to select functions; it may also include a display screen and / or indicator lights and / or speakers, which display the current mode of the machine or the function selection items to the user; it may also include a mobile client program. For the path navigation type automatic cleaning robot 10, the mobile client can show the user a map of the environment where the equipment is located, as well as the location of the machine, and can provide the user with more abundant and humanized function items. Specifically, the cleaning robot has multiple modes, such as working mode, self-cleaning mode, etc. Among them, the working mode refers to the mode in which the cleaning robot performs automatic cleaning operations, and the self-cleaning mode refers to the mode in which the cleaning robot removes dirt from the roller brush and side brush 152 on the base, and automatically collects dirt, and / or automatically washes and dries the mop.
[0052] The cleaning system 150 may be a dry cleaning system 151 and / or a wet cleaning system 153 .
[0053] like Figure 3As shown, the dry cleaning system 151 provided in the embodiment of the present disclosure may include a roller brush, a dust box, a fan, and an air outlet. The roller brush that has a certain interference with the ground sweeps up the garbage on the ground and rolls it to the front of the dust suction port between the roller brush and the dust box, and then the suction gas generated by the fan and passing through the dust box is sucked into the dust box. The dry cleaning system 151 may also include a side brush 152 with a rotating shaft, and the rotating shaft is at a certain angle relative to the ground to move debris into the roller brush area of the cleaning system 150.
[0054] like Figure 3 and Figure 4 As shown, the wet cleaning system 153 provided in the embodiment of the present disclosure may include: a cleaning head 1531, a driving unit 1532, a water supply mechanism, a liquid storage tank, etc. Among them, the cleaning head 1531 can be arranged below the liquid storage tank, and the cleaning liquid inside the liquid storage tank is transmitted to the cleaning head 1531 through the water supply mechanism, so that the cleaning head 1531 performs wet cleaning on the surface to be cleaned. In other embodiments of the present disclosure, the cleaning liquid inside the liquid storage tank can also be directly sprayed onto the surface to be cleaned, and the cleaning head 1531 cleans the surface by evenly applying the cleaning liquid.
[0055] The cleaning head 1531 is used to clean the surface to be cleaned, and the driving unit 1532 is used to drive the cleaning head 1531 to reciprocate along the target surface, which is a part of the surface to be cleaned. The cleaning head 1531 reciprocates along the surface to be cleaned, and a mop is provided on the contact surface between the cleaning head 1531 and the surface to be cleaned. The mop of the cleaning head 1531 is driven by the driving unit 1532 to reciprocate and generate high-frequency friction with the surface to be cleaned, thereby removing stains on the surface to be cleaned; or the mop can be arranged to float and always keep in contact with the cleaning surface during the cleaning process, without the need for the driving unit 1532 to drive its reciprocating motion.
[0056] like Figure 4 As shown, the driving unit 1532 may also include a driving platform 1533 and a supporting platform 1534. The driving platform 1533 is connected to the bottom surface of the machine body 110 for providing driving force. The supporting platform 1534 is detachably connected to the driving platform 1533 for supporting the cleaning head 1531 and can be lifted and lowered under the drive of the driving platform 1533.
[0057] The wet cleaning system 153 can be connected to the machine body 110 through an active lifting module. When the wet cleaning system 153 is not temporarily involved in the work, for example, the cleaning robot 10 stops at the base station to clean the cleaning head 1531 of the wet cleaning system 153 and fill the liquid storage tank with water; or when encountering a surface to be cleaned that cannot be cleaned by the wet cleaning system 153, the wet cleaning system 153 is raised through the active lifting module.
[0058] Figure 5 FIG. 1 is a flow chart of a clustering method for a three-dimensional map of a cleaning robot according to an exemplary embodiment. Figure 5 The method shown can be applied to the server side of the above system, or to the terminal device of the above system. Specifically, the method includes the following steps:
[0059] Step S501: Obtain a three-dimensional map of the cleaning robot and at least one clustering type.
[0060] The process of constructing a three-dimensional map is as follows: first, obtain a dense point cloud or geometric primitives of the working space through the sensing system of the cleaning robot, then generate a 3D mesh model of the space based on the point cloud or geometric primitives, and then based on the generated mesh model, complete the model mapping, that is, map the texture to the 3DMesh model of the space through coordinates to obtain a three-dimensional map.
[0061] The construction of the three-dimensional map may be performed in advance, and the obtained map may be stored in a pre-terminal device or a terminal device, etc. In addition, the construction of the three-dimensional map may be performed simultaneously with the subsequent construction of the two-dimensional map, which is not limited in this embodiment.
[0062] Step S502: Based on the clustering type, clustering algorithms are used to cluster the point clouds or geometric primitives of the three-dimensional map to obtain a clustered three-dimensional map.
[0063] The selection of clustering type can be freely selected by the user on the terminal device according to his own needs. The clustering type can be one or more, and this embodiment does not make a strict limitation.
[0064] Step S503: Displaying the clustered three-dimensional map.
[0065] In this embodiment, the three-dimensional map can be clustered according to different clustering types and then displayed, which improves the diversity of the three-dimensional map display and meets the viewing needs of users.
[0066] In one implementation, Figure 6 As shown, at least one clustering type includes a face clustering type, and step S502 specifically includes:
[0067] Step S601: Based on the face clustering type, clustering algorithms are used to cluster point clouds or geometric primitives in the three-dimensional map to obtain planes and curved surfaces in the three-dimensional map.
[0068] Among them, the face clustering type refers to clustering point clouds or geometric primitives belonging to planes and curved surfaces in the three-dimensional map, so that after clustering the point clouds or geometric primitives of the three-dimensional map, the planes and curved surfaces in the three-dimensional map can be obtained.
[0069] Step S602: Based on the smoothness of the planes and curved surfaces, cluster the planes or curved surfaces using a clustering algorithm to determine planes and / or curved surfaces belonging to the same category.
[0070] Specifically, based on the different smoothness of each plane and curved surface, the planes or curved surfaces with similar smoothness are clustered through a clustering algorithm, so as to obtain planes and / or curved surfaces belonging to the same category, that is, planes and / or curved surfaces belonging to the same area (such as the same room), or planes and / or curved surfaces belonging to the same object (such as a round table, bed, etc.), so as to accurately construct each area on the three-dimensional map and the objects within the area.
[0071] Step S603: marking the planes and / or curved surfaces of the same type in the three-dimensional map according to a preset marking method to obtain a clustered three-dimensional map.
[0072] It is understandable that the preset marking method can be set by the user or by the staff before the machine leaves the factory, and this embodiment does not strictly limit it. The planes and smooth surfaces obtained by clustering are marked so that the customer can clearly see the clustering results.
[0073] In one implementation, Figure 7 As shown, at least one clustering type includes a regional clustering type, and step S502 specifically includes:
[0074] Step S701: Based on the region clustering type, clustering algorithms are used to cluster point clouds or geometric primitives in the three-dimensional map to obtain at least one target region on the three-dimensional map, where the target region is a region of interest to the user.
[0075] The regional clustering type refers to clustering point clouds or geometric point clouds belonging to the target area on the three-dimensional map. The target area refers to the area that the user is interested in, such as a bedroom, a living room, etc., or an area below a certain height. The number of target areas can be set by the user according to needs, and this embodiment does not strictly limit it.
[0076] For example, the user wants to focus on the bedroom, so the point cloud or geometric primitives belonging to the bedroom in the three-dimensional map can be clustered to obtain the point cloud and geometric primitives belonging to the bedroom area on the three-dimensional map, so that subsequent users can understand the layout of the bedroom and the situation of each item more intuitively and clearly.
[0077] In another embodiment, if the user wants to view the situation in a lower area to understand the working environment of the cleaning robot, the user can set a height, for example, a height of 20 mm, so that the cloud points or geometric primitives with a height equal to or lower than 20 mm in the three-dimensional map can be clustered to obtain cloud points or geometric primitives with a height equal to or lower than 20 mm on the three-dimensional map, so that subsequent users can understand the working environment of the cleaning robot and the layout of obstacles more intuitively and clearly.
[0078] Step S702: Mark each target area in the three-dimensional map according to a preset marking method to obtain a clustered three-dimensional map.
[0079] Each clustered area is marked so that the user can clearly see the actual situation of the target area to meet the user's needs.
[0080] In this embodiment, the point cloud or geometric primitives of the area of interest to the user can be clustered to obtain the clustering results of the area of interest to the user, so that the user can understand the actual situation of the area of interest, thereby improving the diversity of three-dimensional map display and meeting customer needs.
[0081] In one implementation, Figure 8 As shown, at least one clustering type includes a semantic clustering type, and step S502 specifically includes:
[0082] Step S801: Based on the semantic clustering type, clustering algorithms are used to cluster point clouds or geometric primitives in the three-dimensional map to obtain semantic information of the three-dimensional map.
[0083] The semantic clustering type refers to clustering based on the semantic information of the three-dimensional map. The semantic information includes but is not limited to the type, shape, color and name of objects in all or part of the three-dimensional map. For example, the user needs to view all objects on the three-dimensional map, so that different types of objects and objects of the same type on the three-dimensional map can be clustered through semantic clustering, so that the semantic information on the three-dimensional map can be obtained, and the real situation of each object on the three-dimensional map can be understood.
[0084] Step S802: Mark each piece of semantic information in the three-dimensional map according to a preset marking method to obtain a clustered three-dimensional map.
[0085] Each clustered area is marked so that the user can clearly see the real situation of each object to meet the user's needs.
[0086] In this embodiment, semantic information of the three-dimensional map can be obtained through semantic clustering, so that users can fully and truly understand the actual situation of each body of the three-dimensional map, thereby improving the diversity of three-dimensional map display and meeting customer needs.
[0087] Furthermore, in the above embodiments, the clustering algorithms involved include K-means algorithm, GMM algorithm, DBSCAN algorithm, spectral clustering algorithm or semantic clustering algorithm of deep learning.
[0088] Specifically, the K-means algorithm: initialize multiple cluster centers, classify points according to their distances from the cluster centers, and recalculate each cluster center after classification. This is repeated, and point cloud segmentation is completed after convergence.
[0089] GMM algorithm: Gaussian mixture model clustering, this method regards each class as a Gaussian distribution in space, uses the points belonging to the class to calculate its mean and variance, and uses the class density function to classify the points based on the mean and variance of each class. Given the initial Gaussian distribution mean and variance, it is iteratively solved, and the point cloud segmentation is completed after convergence.
[0090] Spectral clustering algorithm: Its main idea is to regard all data as points in space, which can be connected by edges. The edge weight between two points that are far away is low, while the edge weight between two points that are close is high. By cutting the graph composed of all data points, the sum of the edge weights between different subgraphs after cutting is as low as possible, while the sum of the edge weights within the subgraph is as high as possible, so as to achieve the purpose of clustering.
[0091] DBSCAN algorithm: (Density-Based Spatial Clustering of Applications with Noise) is a density-based spatial clustering algorithm. The algorithm divides areas with sufficient density into clusters and finds clusters of arbitrary shapes in a spatial database with noise. It defines a cluster as the largest set of density-connected points. The density is determined by the cluster radius, that is, points with adjacent distances less than a certain cluster radius are classified into one class. When the number of points in a class is less than the specified parameter, the points in the class are considered outliers as a whole.
[0092] Deep learning semantic clustering algorithms include but are not limited to PointNet++, PointSIFT, SGPN, SPG and other algorithms. This type of algorithm requires the construction of a training set for the application scenario, including point clouds and their labels. The network's internal parameters are trained by the gradient descent method so that the network has the ability to semantically segment actual data. At the same time, the instance information of the training set is annotated, which can answer the instance segmentation of the point cloud - that is, it can distinguish different individuals under the same category.
[0093] Furthermore, the clustering results of geometric primitives, the former of which can be formed by voting after clustering the point cloud.
[0094] Furthermore, in the above embodiment, the preset marking method includes color marking.
[0095] By color marking, the clustering results are made different in color from other parts on the three-dimensional map, so that the clustering results can be displayed more clearly, so that users can see the clustering results more directly.
[0096] Furthermore, in the above embodiment, after step S503, the following steps are further included:
[0097] Stores cluster types and corresponding clustered 3D maps.
[0098] The clustering type and the corresponding clustered 3D map are stored, so that after obtaining the same clustering type, the corresponding clustered 3D map can be directly retrieved without re-clustering. In addition, the user can also selectively view the previously stored clustering type and the corresponding clustered 3D map.
[0099] Second, as Fig. 9 As shown, an embodiment of the present invention provides a clustering device for a three-dimensional map of a cleaning robot, comprising:
[0100] An acquisition module 901 is used to acquire a three-dimensional map of the cleaning robot and at least one clustering type;
[0101] A clustering module 902 is used to cluster the point cloud or geometric primitives of the three-dimensional map using a clustering algorithm based on the clustering type to obtain a clustered three-dimensional map;
[0102] The display module 903 is used to display the clustered three-dimensional map.
[0103] In a third aspect, an embodiment of the present invention provides an electronic device, including a processor and a memory, wherein the memory is used to store at least one executable instruction, and the executable instruction enables the processor to execute the steps of the above-mentioned clustering method of the three-dimensional map of the cleaning robot.
[0104] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium storing computer program instructions, which, when called and executed by a processor, implement the steps of the above-mentioned method for clustering the three-dimensional map of the cleaning robot.
[0105] The processor may be a central processing unit (CPU), or an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present invention. The one or more processors included in the computer device may be processors of the same type, such as one or more CPUs; or processors of different types, such as one or more CPUs and one or more ASICs.
[0106] The memory is used to store programs. The memory may include a high-speed RAM memory, and may also include a non-volatile memory, such as at least one disk memory.
[0107] The computer-readable storage medium may be a ROM, a random access memory (RAM), a CD-ROM, a magnetic tape, a floppy disk, an optical data storage device, and the like.
[0108] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium storing computer program instructions, which, when called and executed by a processor, perform the steps of the above-mentioned method for processing a three-dimensional map of a cleaning robot.
[0109] The present invention has been described by the above embodiments, but it should be understood that the above embodiments are only for the purpose of example and description, and are not intended to limit the present invention to the scope of the described embodiments. In addition, it can be understood by those skilled in the art that the present invention is not limited to the above embodiments, and more variations and modifications can be made according to the teachings of the present invention, and these variations and modifications all fall within the scope of the protection claimed by the present invention. The protection scope of the present invention is defined by the attached claims and their equivalents.
Claims
1. A clustering method for a three-dimensional map of a cleaning robot, characterized in that: include: Obtaining a three-dimensional map of the cleaning robot and at least one clustering type; Based on the clustering type, clustering the point cloud or geometric primitives of the three-dimensional map using a clustering algorithm to obtain a clustered three-dimensional map; The clustered three-dimensional map is displayed.
2. The method according to claim 1, characterized in that The at least one clustering type includes a face clustering type, and based on the clustering type, clustering the point cloud or geometric primitives of the three-dimensional map using a clustering algorithm to obtain a clustered three-dimensional map includes: Based on the face clustering type, clustering the point cloud or geometric primitives in the three-dimensional map using a clustering algorithm to obtain planes and curved surfaces in the three-dimensional map; Based on the smoothness of the planes and the curved surfaces, clustering the planes and the curved surfaces using the clustering algorithm to determine planes and / or curved surfaces belonging to the same category; In the three-dimensional map, the planes and / or the curved surfaces of the same type are marked according to a preset marking method to obtain a clustered three-dimensional map.
3. The method according to claim 1, characterized in that The at least one clustering type includes a regional clustering type, and based on the clustering type, clustering the point cloud or geometric primitives of the three-dimensional map using a clustering algorithm to obtain a clustered three-dimensional map includes: Based on the region clustering type, clustering the point cloud or geometric primitives in the three-dimensional map using a clustering algorithm to obtain at least one target region on the three-dimensional map, wherein the target region is a region of interest to the user; Each target area in the three-dimensional map is marked according to a preset marking method to obtain a clustered three-dimensional map.
4. The method according to claim 1, characterized in that The at least one clustering type includes a semantic clustering type, and based on the clustering type, clustering the point cloud or geometric primitives of the three-dimensional map using a clustering algorithm to obtain a clustered three-dimensional map includes: Based on the semantic clustering type, clustering the point cloud or geometric primitives in the three-dimensional map using a clustering algorithm to obtain semantic information of the three-dimensional map; In the three-dimensional map, each piece of semantic information is marked according to a preset marking method to obtain a clustered three-dimensional map.
5. The method according to claim 1, characterized in that The clustering algorithm includes K-means algorithm, GMM algorithm, DBSCAN algorithm, spectral clustering algorithm or semantic clustering algorithm of deep learning.
6. The method according to any one of claims 2 to 4, characterized in that: The preset marking method includes color marking.
7. The method according to claim 1, characterized in that After displaying the clustered three-dimensional map, the method further includes: The clustering type and the corresponding clustered three-dimensional map are stored.
8. A clustering device for a three-dimensional map of a cleaning robot, characterized in that: include: An acquisition module, used to acquire a three-dimensional map of the cleaning robot and at least one clustering type; A clustering module, used for clustering the point cloud or geometric primitives of the three-dimensional map using a clustering algorithm based on the clustering type to obtain a clustered three-dimensional map; The display module is used to display the clustered three-dimensional map.
9. An electronic device, characterized in that: It comprises a processor and a memory, wherein the memory is used to store at least one executable instruction, and the executable instruction enables the processor to execute the steps of the clustering method of the three-dimensional map of the cleaning robot as described in any one of claims 1-7.
10. A computer-readable storage medium, characterized in that: Computer program instructions are stored, and when the computer program instructions are called and executed by a processor, the steps of the clustering method of the three-dimensional map of the cleaning robot described in any one of claims 1-7 are implemented.
Citation Information
Patent Citations
Improved multichannel spectral clustering algorithm-based cleaning robot map segmentation method
CN107316308A
Clean area extraction method and device, equipment and storage medium
CN116841286A
Method and apparatus for selecting clusterings to classify a predetermined data set
US20120197888A1
Attribute based map marker clustering
US20150364118A1
