Cleaning path planning method and cleaning robot

Abstract

The application discloses a cleaning path planning method, which comprises the following steps: acquiring top vision image data, wherein the top vision image data is used to represent spatial image data in a plane parallel to and above a plane where a path to be planned is located; identifying a visual occlusion area and a visual non-occlusion area in the top vision image data; determining a minimum circumscribed polygon of the visual occlusion area based on the identified visual occlusion area; selecting one side with the longest length of the minimum circumscribed polygon as an observation reference side for guiding a coverage path in a visual non-occlusion area adjacent to the minimum circumscribed polygon; and planning a cleaning path for the visual occlusion area with the observation reference side as an entry and exit side. The application is beneficial to improving cleaning efficiency and accuracy.

Description

Technical Field

[0001] This invention relates to the field of cleaning robots, and in particular, to a cleaning path planning method. Background Technology

[0002] Cleaning path planning is one of the effective ways for cleaning robots to ensure cleaning efficiency and effectiveness.

[0003] Existing cleaning path planning relies on machine vision positioning based on images captured by a front-facing image acquisition device installed in front of the cleaning robot. However, due to the limited field of view of the front-facing image acquisition device, the efficiency and accuracy of cleaning path planning are unsatisfactory. Summary of the Invention

[0004] This invention provides a cleaning path planning method to improve the efficiency and accuracy of cleaning path planning.

[0005] The first aspect of this invention provides a cleaning path planning method, the method comprising:

[0006] Acquire top-view visual image data, which represents the spatial image data located above and parallel to the plane containing the path to be planned.

[0007] Identify visually occluded and unoccluded regions in top-view visual image data.

[0008] Based on the identified visual occlusion region, determine the minimum bounding polygon of the visual occlusion region.

[0009] In the visually unobstructed region adjacent to the minimum bounding polygon, the longest side of the minimum bounding polygon is selected as the observation reference side to guide the coverage path.

[0010] Using the observation reference edge as the ingress and egress edges, a cleaning path is planned for the visually obstructed area.

[0011] As one possible implementation, acquiring the top-view visual image data includes:

[0012] Spatial image data is acquired by a top-mounted image acquisition device located on top of the cleaning robot in a plane above the support surface of the cleaning robot and parallel to the support surface;

[0013] The step of selecting the longest edge of the smallest bounding polygon as the observation reference edge for guiding the coverage path includes:

[0014] One side of the longest side of the minimum bounding polygon is extended by a set distance threshold towards the visually unobstructed area adjacent to the minimum bounding polygon to obtain the extended side, which serves as the observation reference side.

[0015] As one possible implementation, the planning of cleaning paths for the visually occluded area using the observation reference edge as the input and output edges includes:

[0016] Plan a bow-shaped path within the visual occlusion area in a direction perpendicular to the observation reference edge, so that the visual occlusion area is covered by the bow-shaped path. The straight path in the bow-shaped path is perpendicular to the observation reference edge, and the turning path used to connect two adjacent straight paths in the bow-shaped path is parallel to or coincides with the observation reference edge.

[0017] As one possible implementation, planning a bow-shaped path within the visual occlusion area includes:

[0018] If no feasible region is detected in the current straight path, the current straight path is terminated, and the shortest path is selected to turn back to the observation reference edge, where the bow-shaped path is readjusted.

[0019] As one possible implementation, determining the minimum bounding polygon of the visual occlusion region based on the identified visual occlusion region includes:

[0020] Based on the identified visual occlusion area, determine the minimum bounding rectangle of the visual occlusion area;

[0021] The step of selecting the longest edge of the smallest bounding polygon as the observation reference edge for guiding the coverage path includes:

[0022] The two longest sides of the smallest bounding rectangle are used as the two observation reference sides.

[0023] As one possible implementation, planning a bow-shaped path within the visual occlusion area includes:

[0024] Following a direction perpendicular to the first observation reference edge among the two observation reference edges, plan a bow-shaped straight path within the visual occlusion area until the current straight path reaches the second observation reference edge among the two observation reference edges, and then plan a bow-shaped turning path at the second observation reference edge.

[0025] At the endpoint of the planned bow-shaped turning path at the current second observation reference edge, a new bow-shaped straight path is planned within the visual occlusion area in a direction perpendicular to the second observation reference edge, until the current straight path reaches the first observation reference edge, and a bow-shaped turning path is planned at the first observation reference edge.

[0026] Return to the step of planning a bow-shaped straight path in the visual occlusion area according to the direction perpendicular to the first observation reference edge of the two observation reference edges, until the visual occlusion area is covered by the bow-shaped path.

[0027] As one possible implementation, the length of the rotary path is determined based on the maximum size of the cleaning component in the cleaning robot that contacts the area being cleaned.

[0028] As one possible implementation, the planning of the straight path further includes:

[0029] If an infeasible region is detected in the current straight path, the current straight path is terminated, and the shortest path is selected to turn back to the observation reference edge from which the current straight path starts. A bow-shaped turning path is planned at the current observation reference edge. This turning path is parallel to or coincides with the current observation reference edge, and the length of this turning path is adjusted to allow the next straight path starting from the current observation reference edge to bypass the length of the infeasible region in the current straight path.

[0030] Following a direction perpendicular to the current observation reference edge, plan a bow-shaped straight path starting from the current observation reference edge within the visual occlusion area until the next straight path reaches another observation reference edge. At the other observation reference edge, plan a bow-shaped turning path that is parallel to or coincides with the other observation reference edge. The length of the turning path is adjusted to cover the area between the next straight path starting from the other observation reference edge and the next straight path starting from the current observation reference edge, and to reach a length where there is no feasible area in the current straight path.

[0031] A second aspect of this application provides a cleaning robot, which includes a memory and a processor. The memory stores a computer program, and the processor is configured to execute the computer program to implement the cleaning path planning method.

[0032] As one possible implementation, the cleaning robot includes an image acquisition device on its top.

[0033] The cleaning path planning method provided by this invention plans paths for visually occluded areas based on visually occluded areas identified by top-view visual image data. This not only avoids the field-of-view limitations of the front-view image acquisition device but also solves the problem of generating cleaning planning paths when top-view visual positioning fails due to occlusion. It reduces long-term occlusion situations to ensure positioning stability. By using the observation reference edge as the entry and exit edge, it avoids excessive repetitive paths, which helps improve cleaning efficiency. It also reduces continuous visual occlusion during each cleaning path switch, improving positioning stability. As a result, the cleaning robot becomes more adaptable to cleaning operations and effectively reduces missed cleaning areas. Attached Figure Description

[0034] Figure 1This is a flowchart illustrating a cleaning path planning method according to an embodiment of this application.

[0035] Figure 2 This is a schematic diagram of the cleaning path planning in this embodiment.

[0036] Figure 3 This is a schematic diagram of the identified visually occluded and visually unoccluded regions.

[0037] Figure 4 This is a schematic diagram of the minimum bounding rectangle of a visual occlusion region.

[0038] Figure 5 This is a schematic diagram of the generation of the observation reference edge in this embodiment.

[0039] Figure 6 This is a schematic diagram of the bow-shaped path planning in this embodiment.

[0040] Figure 7 This is another schematic diagram of the bow-shaped path planning in this embodiment.

[0041] Figure 8 This is a schematic diagram of a cleaning path planning device according to an embodiment of this application.

[0042] Figure 9 This is another schematic diagram of the cleaning path planning device according to an embodiment of this application. Detailed Implementation

[0043] To make the objectives, technical means, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings.

[0044] The applicant's research found that the cleaning path planning of cleaning robots is based on the visual image data of the bearing surface collected by the front image acquisition device. Since the bearing surface is covered with various obstacles, such as various furniture and household items, this not only leads to fragmented cleaning path planning, resulting in low cleaning efficiency, but also makes it almost impossible to plan cleaning paths for bearing surfaces located under beds, tables, etc., thus making it difficult to clean effectively.

[0045] In view of this, embodiments of this application provide a cleaning path planning method that uses top-mounted visual image data to plan the cleaning path, thereby improving the efficiency and accuracy of cleaning path planning.

[0046] See Figure 1 As shown, Figure 1 This is a flowchart illustrating a cleaning path planning method according to an embodiment of this application.

[0047] The method includes:

[0048] Step 101: Obtain top-view visual image data, wherein the top-view visual image data is used to represent spatial image data located above the plane where the path to be planned is located and parallel to the plane.

[0049] As an example, a top-mounted image acquisition device located on top of the cleaning robot acquires spatial image data in a plane above and parallel to the support surface on which the cleaning robot is located. In this way, the top-mounted visual image data is equivalent to a bottom view of the spatial environment in which the cleaning robot is located. It should be understood that the top-mounted visual image data can be acquired in real time, for example, in real time during the cleaning process, or it can be acquired in non-real time, for example, acquired in advance and stored in the cleaning robot body. This application does not limit this.

[0050] Step 102: Identify the visually occluded and unoccluded areas in the top-view visual image data.

[0051] As an example, feature extraction is performed based on top-view visual image data. The extracted features are then used to identify visually occluded and unoccluded regions within the top-view visual image data.

[0052] As another example, trained artificial intelligence models, such as neural network models, are used to identify visually occluded and unoccluded areas in top-view visual image data.

[0053] Step 103: Based on the identified visual occlusion region, determine the minimum bounding polygon of the visual occlusion region.

[0054] As an example, there may be multiple visual occlusion regions. For each visual occlusion region, determine the minimum bounding polygon of that visual occlusion region.

[0055] To reduce complexity, the smallest bounding polygon can be the smallest bounding rectangle.

[0056] Step 104: In the visually unobstructed region adjacent to the minimum bounding polygon, select one of the longest sides of the minimum bounding polygon as the observation reference side to guide the coverage planning path.

[0057] As an example, one side of the longest side of the minimum bounding polygon is extended by a set distance threshold towards the visually unobstructed area adjacent to the minimum bounding polygon, resulting in an extended side that serves as an observation reference side. The distance threshold can be determined based on the dimensions of the cleaning robot itself.

[0058] The number of observation reference edges can be multiple. For example, when the smallest bounding polygon is the smallest bounding rectangle, the two longest sides of the smallest bounding rectangle are used as observation reference edges.

[0059] Step 105: Using the observation reference edge as the in and out edges, plan the cleaning path for the visually obstructed area.

[0060] As an example, a bow-shaped path is planned in the visual occlusion area along a direction perpendicular to the observation reference edge, so that the visual occlusion area is covered by the bow-shaped path. The straight path in the bow-shaped path is perpendicular to the observation reference edge, and the turning path used to connect two adjacent straight paths in the bow-shaped path is parallel to or coincides with the observation reference edge.

[0061] For example, if the visual occlusion area is a minimum bounding rectangle, a bow-shaped straight path can be planned within the visual occlusion area along a direction perpendicular to the first observation reference edge of the two observation reference edges, until the current straight path reaches the second observation reference edge of the two observation reference edges, and then a bow-shaped turning path is planned at the second observation reference edge.

[0062] At the endpoint of the planned bow-shaped turning path at the current second observation reference edge, a new bow-shaped straight path is planned within the visual occlusion area in a direction perpendicular to the second observation reference edge, until the current straight path reaches the first observation reference edge, and a bow-shaped turning path is planned at the first observation reference edge.

[0063] Return to the step of planning a bow-shaped straight path within the visual occlusion area according to the direction perpendicular to the first observation reference edge of the two observation reference edges, until the visual occlusion area is covered by the bow-shaped path. In this way, the planned path is formed within the two observation reference edges.

[0064] Furthermore, if no feasible region is detected in the current straight path, the current straight path is terminated, and the shortest path is selected to turn back to the observation reference edge, where the bow-shaped path is readjusted.

[0065] For example, in the case where the visual occlusion area is the smallest bounding rectangle, if it is detected that there is no feasible area in the current straight path, the current straight path is terminated, and the shortest path is selected to turn back to the observation reference edge that the current straight path starts from. At the current observation reference edge, a bow-shaped turning path is planned. This turning path is parallel to or coincides with the current observation reference edge, and the length of this turning path is adjusted to allow the next straight path starting from the current observation reference edge to bypass the length of the current straight path where there is no feasible area.

[0066] Following a direction perpendicular to the current observation reference edge, plan a bow-shaped straight path starting from the current observation reference edge within the visual occlusion area until the next straight path reaches another observation reference edge. At the other observation reference edge, plan a bow-shaped turning path that is parallel to or coincides with the other observation reference edge. The length of the turning path is adjusted to cover the area between the next straight path starting from the other observation reference edge and the next straight path starting from the current observation reference edge, and to reach the length where there is no feasible area in the current straight path. In this way, the planned path covers the feasible area within the two observation reference edges.

[0067] The cleaning path planning method provided in this application plans cleaning paths for visually occluded areas in the top-view visual image data, so that occluded areas above the plane where the path to be planned can also be planned. This helps to increase the cleaning range and avoid missing cleaning areas. By observing reference edges to plan the path, the cleaning robot can enter the occluded area in the shortest way, which helps to improve cleaning efficiency.

[0068] To facilitate understanding of the embodiments of this application, the following uses a cleaning robot to plan a cleaning path based on top-mounted visual image data as an example. It should be understood that the embodiments of this application are not limited to this specific top-mounted visual image data, and top-mounted visual image data in any other spatial environment can be applied to this application.

[0069] See Figure 2 As shown, Figure 2 This is a schematic diagram of cleaning path planning in this embodiment. The method includes:

[0070] Step 201: Obtain top-view visual image data.

[0071] As an example, the cleaning robot moves around in the environment beforehand, acquiring top-view visual image data through a top-mounted image acquisition device installed on its top to construct a base map.

[0072] Step 202: Based on the acquired top-view visual image data, identify the visually occluded area and the visually unoccluded area.

[0073] See Figure 3 As shown, Figure 3This diagram illustrates the identified visually occluded and unoccluded areas. Green areas represent visually occluded areas, including three areas whose range exceeds a set threshold. Red and yellow areas represent visually unoccluded areas. For example, an unoccluded area is typically the ceiling, which can be understood as a feasible area without obstacles in the top-view visual image data. Visually occluded areas, due to the ceiling obstructing the view, result in unstable localization during the visual localization process and can be understood as unstable localization areas.

[0074] The identification method can be: based on the top-view visual image data, perform feature extraction, and identify the visually occluded and unoccluded regions in the top-view visual image data according to the extracted features; or it can be using a trained artificial intelligence model, such as a neural network model, to identify the visually occluded and unoccluded regions in the top-view visual image data. This application does not limit this method.

[0075] Step 203: Based on the identified visual occlusion area, determine the minimum bounding rectangle of the visual occlusion area.

[0076] As an example, the pixel position information of the minimum bounding rectangle of the visually occluded region is calculated based on the pixel position information of the identified visually occluded region.

[0077] See Figure 4 As shown, Figure 4 This is a schematic diagram of the minimum bounding rectangle of a visually occluded region. In the diagram, the minimum bounding rectangle of the visually occluded region can be determined using the pixel position information of the four vertices a, b, c, and d of the minimum bounding rectangle.

[0078] Step 204: Generate observation reference edge information based on the minimum bounding rectangle.

[0079] As an example, within the feasible region adjacent to the minimum bounding rectangle, the longest side of the minimum bounding rectangle is selected as the reference. For example, within the feasible region adjacent to the minimum bounding rectangle, the longest side is selected, that is, among the four rectangular sides, the rectangular sides that fall into the occlusion area are not counted, and the longest rectangular side is selected from the remaining rectangular sides.

[0080] Based on the selected baseline, an observation reference edge with the same length as the baseline edge is extended perpendicularly to the feasible region to guide the coverage planning path. The extended distance threshold can be determined based on the dimensions of the cleaning robot body, such as its width. In other words, the selected longest rectangle edge is vertically shifted towards the feasible region by a set distance threshold. Extending the baseline edge improves the reliability of path planning and reduces the possibility of positioning failure.

[0081] See Figure 5 As shown, Figure 5 This is a schematic diagram illustrating the generation of the observation reference edge in this embodiment. In the diagram, the longest rectangular edge within the unobstructed visual area is ab. Based on ab, the observation reference edge a1b1 is obtained by expanding the rectangle.

[0082] The number of observation reference edges can be multiple. Using the two longest edges of the smallest bounding rectangle as references, each edge is extended to its adjacent unoccluded visual region, resulting in two observation reference edges, such as... Figure 7 As shown by the blue line segment.

[0083] Step 205: Using the observation reference edge as the input and output edges, perform path planning for the video occlusion area.

[0084] As an example, a bow-shaped path is planned within the minimum bounding rectangle along a direction perpendicular to the observation reference edge. The straight path in the bow is perpendicular to the observation reference edge, and the turning path used to connect two adjacent straight paths can be parallel to or coincide with the observation reference edge.

[0085] It should be understood that in this embodiment, there is no restriction on the starting position of the bow-shaped path on the observation reference edge. That is, the bow-shaped path can start and end at any point on the observation reference edge, as long as the video occlusion area is covered by the planned path. This helps the cleaning robot enter the visual occlusion area with the shortest path, thereby improving cleaning efficiency.

[0086] See Figure 6 As shown, Figure 6 This is a schematic diagram of a bow-shaped path planning in this embodiment. In the diagram, a bow-shaped straight path is planned starting from position A on the observation reference edge until an obstacle is reached. At the obstacle, a bow-shaped turning path is planned along the edge. This process is repeated until the video occlusion area within the smallest bounding rectangle is covered by the planned path.

[0087] See Figure 7 As shown, Figure 7 This is another schematic diagram of the bow-shaped path planning in this embodiment. Starting from position A of the observation reference edge b1a1, a bow-shaped straight path is planned along the direction perpendicular to the observation reference edge b1a1 until the observation reference edge c1d1 is reached. Then, a turning path is planned at the observation reference edge c1d1. Finally, at the end of the turning path, a straight path is planned along the direction perpendicular to the observation reference edge c1d1.

[0088] If an infeasible region is detected on the current straight path, for example, an obstacle is detected, then the shortest path is used to return to the starting point of the current straight path, such as... Figure 7As shown by the dashed line, a loop path is then planned, which is parallel to or coincides with the current observation reference edge. The length of this loop path is adjusted to ensure that the next straight path starting from the current observation reference edge b1a1 bypasses the length of the current straight path where there is no feasible region. Figure 7 The AB path is then returned to the step of planning a bow-shaped straight path from position A of the observation reference edge b1a1 along a direction perpendicular to the observation reference edge b1a1, until the observation reference edge c1d1 is reached; at the observation reference edge c1d1, a turning path is planned, which is parallel to or coincides with the current observation reference edge c1d1, and the length of the turning path is adjusted to cover the area between the next straight path starting from the current observation reference edge c1d1 and the straight path starting from the observation reference edge b1a1, and reaches the length where there is no feasible area in the current straight path, such as... Figure 7 The CD path in the file.

[0089] Step 206: The cleaning robot moves along the planned path to perform the cleaning operation.

[0090] As an example, the cleaning robot converts the planned path into a planned path in the world coordinate system and moves according to the planned path to perform cleaning operations on the surface on which the cleaning robot is located.

[0091] Step 207: Determine whether an obstacle has been detected during the current movement.

[0092] If so, select the shortest path to return to the observation reference edge, and rotate at the observation reference edge according to the planned rotation path, so as to move according to the next planned straight path.

[0093] Otherwise, return to step 206 until all areas obscured by the video have been cleaned.

[0094] In this embodiment, the planned path is executed after the planned path is completed. It should be understood that the planned path can also be generated and executed in real time, and this embodiment does not limit this.

[0095] See Figure 8 As shown, Figure 8 This is a schematic diagram of a cleaning path planning device according to an embodiment of this application. The device includes:

[0096] The acquisition module is used to acquire top-view visual image data, wherein the top-view visual image data is used to represent spatial image data located above the plane where the path to be planned is located and parallel to the plane.

[0097] The recognition module is used to identify visually occluded areas and visually unoccluded areas in the top-view visual image data.

[0098] The planning module is used to determine the minimum bounding polygon of the visually occluded area based on the identified visually occluded area. In the visually unoccluded area adjacent to the minimum bounding polygon, the longest side of the minimum bounding polygon is selected as the observation reference edge to guide the coverage path. The cleaning path is planned for the visually occluded area using the observation reference edge as the in and out edge.

[0099] See Figure 9 As shown, Figure 9 This is another schematic diagram of a cleaning path planning device according to an embodiment of this application. The device includes a memory and a processor, wherein the memory stores a computer program, and the processor is configured to execute the computer program to implement the steps of the cleaning path planning method according to an embodiment of this application.

[0100] The memory may include random access memory (RAM) or non-volatile memory (NVM), such as at least one disk storage device. Optionally, the memory may also be at least one storage device located remotely from the aforementioned processor.

[0101] The processors mentioned above can be general-purpose processors, including central processing units (CPUs), network processors (NPs), etc.; they can also be digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0102] This invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the cleaning path planning method of this application.

[0103] For the device / network-side equipment / storage medium embodiments, since they are basically similar to the method embodiments, the description is relatively simple, and relevant parts can be referred to in the description of the method embodiments.

[0104] In this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, without necessarily requiring or implying any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0105] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A cleaning path planning method, characterized in that, The method includes: Acquire top-view visual image data, which represents the spatial image data located above and parallel to the plane containing the path to be planned. The system identifies visually occluded and unoccluded regions in the top-view visual image data. The unoccluded region is the ceiling, which represents a feasible area without obstacles in the top-view visual image data. The visually occluded region, due to the ceiling obstructing the view, experiences unstable localization results during the visual localization process and is thus classified as an unstable localization region. Based on the identified visually occluded regions, the minimum bounding polygon of each region is determined. In the visually unobstructed region adjacent to the minimum bounding polygon, the longest side of the minimum bounding polygon is selected as the observation reference side to guide the coverage path. Using the observation reference edge as the ingress and egress edges, a cleaning path is planned for the visually obstructed area.

2. The cleaning path planning method as described in claim 1, characterized in that, The acquisition of top-view visual image data includes: Spatial image data is acquired by a top-mounted image acquisition device located on top of the cleaning robot in a plane above the support surface of the cleaning robot and parallel to the support surface; The step of selecting the longest edge of the smallest bounding polygon as the observation reference edge for guiding the coverage path includes: One side of the longest side of the minimum bounding polygon is extended by a set distance threshold towards the visually unobstructed area adjacent to the minimum bounding polygon to obtain the extended side, which serves as the observation reference side.

3. The cleaning path planning method as described in claim 1 or 2, characterized in that, The planning of cleaning paths for visually obstructed areas, using the observation reference edge as the ingress and egress edges, includes: Plan a bow-shaped path within the visual occlusion area in a direction perpendicular to the observation reference edge, so that the visual occlusion area is covered by the bow-shaped path. The straight path in the bow-shaped path is perpendicular to the observation reference edge, and the turning path used to connect two adjacent straight paths in the bow-shaped path is parallel to or coincides with the observation reference edge.

4. The cleaning path planning method as described in claim 3, characterized in that, Planning a bow-shaped path within the visually obstructed area includes: If no feasible region is detected in the current straight path, the current straight path is terminated, and the shortest path is selected to turn back to the observation reference edge, where the bow-shaped path is readjusted.

5. The cleaning path planning method as described in claim 4, characterized in that, The step of determining the minimum bounding polygon of the visual occlusion region based on the identified visual occlusion region includes: Based on the identified visual occlusion area, determine the minimum bounding rectangle of the visual occlusion area; The step of selecting the longest edge of the smallest bounding polygon as the observation reference edge for guiding the coverage path includes: The two longest sides of the smallest bounding rectangle are used as the two observation reference sides.

6. The cleaning path planning method as described in claim 5, characterized in that, Planning a bow-shaped path within the visually obstructed area includes: Following a direction perpendicular to the first observation reference edge among the two observation reference edges, plan a bow-shaped straight path within the visual occlusion area until the current straight path reaches the second observation reference edge among the two observation reference edges, and then plan a bow-shaped turning path at the second observation reference edge. At the endpoint of the planned bow-shaped turning path at the current second observation reference edge, following a direction perpendicular to the second observation reference edge, plan the next bow-shaped straight path within the visual occlusion area until the current straight path reaches the first observation reference edge, and then plan a bow-shaped turning path at the first observation reference edge. Return to the step of planning a bow-shaped straight path in the visual occlusion area according to the direction perpendicular to the first observation reference edge of the two observation reference edges, until the visual occlusion area is covered by the bow-shaped path.

7. The cleaning path planning method as described in claim 6, characterized in that, The length of the rotation path is determined based on the maximum size of the cleaning component in the cleaning robot that comes into contact with the area being cleaned.

8. The cleaning path planning method as described in claim 6, characterized in that, The planning of the straight path further includes: If an infeasible region is detected in the current straight path, the current straight path is terminated, and the shortest path is selected to turn back to the observation reference edge from which the current straight path starts. A bow-shaped turning path is planned at the current observation reference edge. This turning path is parallel to or coincides with the current observation reference edge, and the length of this turning path is adjusted to allow the next straight path starting from the current observation reference edge to bypass the length of the infeasible region in the current straight path. Following a direction perpendicular to the current observation reference edge, plan a bow-shaped straight path starting from the current observation reference edge within the visual occlusion area until the next straight path reaches another observation reference edge. At the other observation reference edge, plan a bow-shaped turning path that is parallel to or coincides with the other observation reference edge. The length of the turning path is adjusted to cover the area between the next straight path starting from the other observation reference edge and the next straight path starting from the current observation reference edge, and to reach a length where there is no feasible area in the current straight path.

9. A cleaning robot, characterized in that, The cleaning robot includes a memory and a processor, the memory storing a computer program, and the processor being configured to execute the computer program to implement the cleaning path planning method according to any one of claims 1 to 8.

10. The cleaning robot as described in claim 9, characterized in that, The cleaning robot has an image acquisition device on its top.

Images