Automatic working method and self-moving equipment
By segmenting and expanding the images collected from mobile devices, non-passable areas are calibrated, which solves the safety hazards of the equipment when identifying boundaries, ensures that it drives and works within the actual physical boundaries, and avoids driving out of the boundaries and other Security risks.
Patent Information
- Application Number
- CN202410503345.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Priority Date
- 2023-09-01
- Filing Date
- 2024-04-25
- Publication Date
- 2024-12-03
AI Technical Summary
Since mobile devices bury cables at the boundary to identify the boundary line for re-charging or looking for transition signs, there are safety hazards, such as driving out of the boundary or damage to the camera, especially in the context of intelligent development of visual mobile devices.
By collecting images from the front of the mobile device, segmenting the images to identify target non-grass areas, such as fences or fence areas, and performing out-expanding processing, calibrating the out-expanding area as a non-passable area, and controlling the driving of the equipment according to the calibration results and Work to match the actual physical boundaries.
It effectively avoids safety issues such as driving out of the boundary from mobile devices, ensures that the equipment works within the actual physical boundaries, and improves safety and accuracy.
Smart Images

Figure CN119065362A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of equipment automation control, and in particular to an automatic working method and a self-moving equipment. Background Art
[0002] At present, autonomous vehicles usually need to bury cables at the boundaries as identifiable boundaries of the autonomous vehicles, and search for these boundaries to complete the return to charge or find transition signs. With the advent of the intelligent era, autonomous vehicles that rely on visual movement have become a market development trend. Summary of the invention
[0003] In order to overcome the defects of the prior art, the problem to be solved by the present invention is to provide an automatic working method and a self-moving device that can work safely in a working area.
[0004] The technical solution adopted by the present invention to solve the problems of the prior art is: an automatic working method, the method is applied to a self-moving device, the self-moving device travels and / or works in a work area with limited boundaries, the method comprising:
[0005] Capturing the image in front of the mobile device;
[0006] Segmenting the image to obtain a segmentation result, wherein the segmentation result includes a target non-grass area;
[0007] Expanding the target non-grass area, and marking the area determined after the expansion as a non-passable area, wherein the non-passable area is an area where the self-moving device cannot travel and / or work;
[0008] The driving and / or operation of the self-moving device is controlled according to the calibration result.
[0009] In one embodiment of the present application, the target non-grass area includes at least one of the following: a fence area, a fence area, and there is a gap in the vertical direction between the fence or a part of the bottom of the fence and the boundary.
[0010] In one embodiment of the present application, expanding the target non-grass area includes:
[0011] The target non-grass area is image-expanded according to a preset rule to obtain an expanded first area, and the first area is marked as a non-passable area.
[0012] In one embodiment of the present application, the preset rules include: expanding the target non-grass area up and down and / or left and right.
[0013] In one embodiment of the present application, the preset rule includes: expanding the target non-grass area up and down and expanding the target non-grass area left and right, wherein the pixel points of the up and down expansion and the pixel points of the left and right expansion are different.
[0014] In one embodiment of the present application, expanding the target non-grass area includes:
[0015] A circumscribed polygon of the target non-grass area is determined, and the circumscribed polygon is marked as the non-passable area.
[0016] In one embodiment of the present application, before determining the circumscribed polygon of the target non-grass area, the method further includes:
[0017] When at least two target non-grass areas are identified, the target non-grass area with the largest area and the target non-grass area that meets the requirements are merged, and the merged area is marked as the non-passable area. Among the target non-grass areas that meet the requirements, there are pixels located in the lower half of the image and in the lower right corner of the image.
[0018] In one embodiment of the present application, the target non-grass area that meets the requirements has a pixel point with a horizontal coordinate of 511 and a vertical coordinate greater than or equal to 256 in the image coordinate system.
[0019] In one embodiment of the present application, after determining the circumscribed polygon of the target non-grass area, the method further includes:
[0020] The circumscribed polygon is enlarged by the same distance to obtain an enlarged second area, and the second area is marked as a non-passable area, and the distance is related to the suspended height of the target non-grass area.
[0021] In one embodiment of the present application, the distance is set by a user.
[0022] An embodiment of the present invention further provides a self-moving device, the self-moving device comprising:
[0023] A visual sensor, the visual sensor is used to collect the image in front of the mobile device;
[0024] A controller is electrically connected to the visual sensor, and the controller segments the image to obtain a segmentation result, wherein the segmentation result includes a target non-grass area; the target non-grass area is expanded, and the area determined after the expansion is calibrated as a non-passable area, wherein the non-passable area is an area where the self-moving device cannot travel and / or work, and the travel and / or work of the self-moving device is controlled according to the calibration result.
[0025] Compared with the prior art, the beneficial effect of the present invention is that by expanding the divided fence and other areas, the boundaries of the expanded fence area are matched with the physical boundaries determined by the fence in the actual scene. Therefore, when the lawn mower travels according to the image processed by the automatic working method, it will travel in the working area determined by the actual physical boundaries, thereby avoiding safety problems such as the lawn mower driving out of the boundary. BRIEF DESCRIPTION OF THE DRAWINGS
[0026] The above-mentioned objectives, technical solutions and beneficial effects of the present invention can be achieved through the following drawings:
[0027] Figure 1 is a schematic diagram of an automatic working system according to an embodiment of the present invention;
[0028] Figure 2 is a schematic diagram of the expansion of the target non-grass area in one embodiment of the present invention;
[0029] Figure 3 is a schematic diagram of merging multiple target non-grass areas in one embodiment of the present invention;
[0030] Figure 4 is a schematic diagram of an automatic working method of a lawn mower according to an embodiment of the present invention;
[0031] Figure 5 It is a schematic diagram of the expansion of the target non-grass area in another embodiment of the present invention. DETAILED DESCRIPTION
[0032] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0033] The terms used in the specification of the present invention herein are only for the purpose of describing specific embodiments and are not intended to limit the present invention. In addition, the term "and / or" used herein includes any and all combinations of one or more related listed items.
[0034] like Figure 1The automatic working system of this specific embodiment is shown, which may include: a self-moving device 1, a boundary 7, and a base station 3. Among them, the self-moving device 1 travels and works in the working area 4 defined by the boundary 7, and the base station 3 can be used to return and replenish energy for the self-moving device when the energy is insufficient. The boundary 7 can be the periphery of the entire working area, called the outer boundary, which is usually connected end to end to close the working area 4, and can be electronic or physical. The physical boundary can be simply a natural physical boundary formed by the junction between the working area 4 and the non-working area, such as: a natural boundary between grass and non-grass, or a boundary formed by walls, fences, railings, etc. For example, when the user is adjacent to the neighbor's lawn, the user can use a method such as Figure 1 The fence 2 shown forms a boundary. The electronic boundary can be formed by laying wires around the working area 4 and using a virtual boundary signal, such as an electromagnetic signal, an acoustic signal or a light signal, etc., emitted by a boundary signal generating device connected to the wires. There may also be an area 5 in the working area that is not suitable for the self-moving device 1 to work, and the area 5 forms a boundary, such as a flower bed, a pool, an obstacle, etc., which can be called an inner boundary, and the part outside the inner boundary is the working area.
[0035] The self-moving device 1 can be an automatic lawn mower, a sweeping robot, an automatic snow sweeper, or other unattended devices, which automatically travel on the surface of the work area to mow the grass, vacuum, or sweep snow. Of course, the self-moving device is not limited to an automatic lawn mower, a sweeping robot, or an automatic snow sweeper, and can also be other unattended devices, which is not limited in this application.
[0036] In the following specific embodiments, a detailed description is given by taking the automatic working system as an automatic lawn mower system, that is, the self-mobile device 1 as an automatic lawn mower as an example.
[0037] The automatic lawn mower includes a shell, a driving module, a working module, a controller, and an energy module. The controller connects and controls the driving module and the working module to realize the automatic driving and working of the automatic lawn mower. In the embodiment of the present application, the automatic lawn mower may also include a visual sensor 11 for collecting images. After collecting the image, the lawn mower performs image segmentation on the grass, soil, fallen leaves, fences, etc., and controls the lawn mower to drive in the segmented grass or other categories (such as soil) that the user allows the lawn mower to drive. The lawn mower can mow the grass inside the working area by randomly driving. After completing the internal mowing, it can also cut the grass on the boundary or return to charge by driving along the edge.
[0038] like Figure 1 As shown in the figure, when the lawns of two households are adjacent, they usually use fences, railings or hedges with a partially suspended bottom to divide the area of their own lawns. In this case, the fences, railings or hedges are part of the boundary. Figure 1In this scenario, when the visual lawn mower is mowing the user's lawn, it can use the fence position 2 in the captured image as the boundary between grass and non-grass, and control its movement and work in the working area according to the boundary. That is, the lawn mower will continue to move according to the image display result until it reaches the position corresponding to the boundary of the fence area in the image. Since the suspended part of the fence is displayed as grass in the image, corresponding to the actual scene, after the lawn mower reaches the suspended part of the fence boundary, it will continue to move forward and go out of bounds (such as Figure 1 8 is the projection of the fence on the ground, that is, the actual physical boundary. After the lawn mower reaches the physical boundary 8, it will continue to move forward), that is, the lawn mower may go out of the boundary, or the camera may be damaged in the process of continuing to move forward, or the lawn mower may be stuck in the fence and other safety issues.
[0039] In view of the above problems, the present application proposes an automatic working method, in which the fence and other areas after segmentation are expanded so that the boundary of the expanded fence area matches the physical boundary determined by the fence in the actual scene. When the lawn mower drives according to the image processed by the automatic working method, it will drive in the working area determined by the actual physical boundary, thereby avoiding safety problems such as the lawn mower driving out of the boundary. Specifically, Figure 4 As shown, the automatic working method includes the following steps:
[0040] S401: Capturing the image in front of the mobile device.
[0041] S402: Segment the image to obtain a segmentation result, where the segmentation result includes a target non-grass area.
[0042] In the embodiment of the present application, the target non-grass area includes a fence area, a fence area, and other non-grass areas in the image that need to be further optimized. There is a vertical distance (such as 1 / 200 m) between the fence 2 or a portion of the bottom of the fence and the physical boundary. Figure 1 There is a gap between the bottom of the fence 2 in the image and the physical boundary 8 corresponding to the fence projection), that is, the bottom of the fence or the fence is partially suspended. Directly using the image to perform mowing operations will cause the lawn mower to drive to a non-passable area, which may cause safety hazards such as the lawn mower going out of bounds or the camera being damaged.
[0043] In one embodiment of the present application, the lawn mower collects images through a visual sensor during driving, and performs segmentation processing on the collected images to identify grass areas, stone slab areas, fences and other areas in the images (such as Figure 2 The area enclosed by the line segment 21 in the image is the captured image area), wherein the fence area is the target non-grass area, and the target non-grass area is expanded (such as Figure 2The area enclosed by the line segment 81 in the image is the captured image area), and the expanded area is marked as a non-passable area (a non-passable area is an area where the lawn mower is not allowed to travel or mow), and the boundary 81 of the non-passable area is Figure 1 The dotted line 8 in the figure corresponds to the recognition result, and the lawn mower is controlled to perform mowing work in the passable area.
[0044] S403: Expand the target non-grass area, and mark the area determined after the expansion as a non-passable area, where the non-passable area is an area where the self-moving device cannot travel and / or work.
[0045] In an embodiment of the present application, after the target non-grass area is identified, the target non-grass area can be expanded, and the area determined after the expansion can be marked as a non-passable area, and the non-passable area is an area where the self-moving device cannot travel and / or work. By expanding the target non-grass area, that is, stretching the target non-grass area, the boundary of the fence area after the expansion matches the physical boundary determined by the fence in the actual scene, so that the lawn mower can travel in the working area determined by the actual physical boundary.
[0046] Before expanding the target non-grass area, the outer contour of the target non-grass area can be identified first, and whether there is a part exceeding the preset threshold in the outer contour, and the part exceeding the preset threshold is smoothed or removed according to other contours to obtain a relatively flat target non-grass area. Corresponding to the scene where the target non-grass area is a fence, such as Figure 1 as well as Figure 5 As shown in the figure, since there are support columns 6 and the like in the fence, there is a protrusion 25 in the fence area obtained by segmenting the image taken while the lawn mower is moving along the edge. Since the support column 6 is not suspended, if the protrusion 25 is also expanded, the machine will inevitably miss the grass area when walking along the boundary after the expansion process. Therefore, the protrusion 25 can be removed by image processing to obtain the removed part 26, and then the part without the protrusion 25 can be expanded to obtain Figure 2 The expanded portion 81 in the middle is used to avoid unnecessary missing cutting.
[0047] Specifically, in one embodiment of the present application, the fence area boundary can be expanded in the following manner, including: performing image expansion on the target non-grass area according to a preset rule to obtain a first area after expansion, and marking the first area as a non-passable area. In one embodiment, the target non-grass area can be expanded horizontally, vertically, or in all directions. When performing all-direction expansion, the target non-grass area can be expanded in the same proportion up, down, left, and right (square expansion), or in different proportions (rectangular expansion or elliptical expansion, etc.), or horizontal expansion or vertical expansion can be performed separately, and the pixels during horizontal expansion or vertical expansion are different. Among them, horizontal expansion is mainly for image processing made in the length direction of the target non-grass area in the image, and vertical expansion is mainly for image processing made in the height direction of the target non-grass area. For example, each pixel point in the fence area is image expanded, and 30 pixels are expanded vertically and 3 pixels are expanded horizontally during image expansion. The inflection point of the contour of the expanded image is smoother, and the contour inflection point is approximately elliptical. In another embodiment, the height direction of the target non-grass area may be detected first, and vertical expansion is performed only in the height direction, while no expansion or a smaller expansion is performed in the horizontal direction.
[0048] On the one hand, there will be a certain degree of damage when processing the collected image, which makes it unable to reflect the real boundary situation in the horizontal direction and in the vertical direction. On the other hand, since the fence and the like are suspended in the air, the collected image corresponds to grass in the vertical direction, resulting in the boundary not corresponding to the real physical boundary. In view of these two situations, in the embodiments of the present application, preferably, an image expansion of the target non-grass area is performed in an approximately elliptical shape, and during the image expansion, the horizontal (left and right) direction expands less and the vertical (up and down) direction expands more. Specifically, the boundary obtained by the left and right expansion is closer to the real boundary length in length, and the area obtained by the up and down expansion is closer to the real fence area. Further, the area boundary obtained by the up and down expansion of more pixels is closer to the projection of the real fence area on the grass. The image obtained after expansion is used as a non-passable area, so that when the visual lawn mower identifies the corresponding boundary of the non-passable area during operation, it executes corresponding strategies such as retreat and turn.
[0049] Specifically, in another embodiment of the present application, the fence area boundary can be expanded in the following manner, including: determining a circumscribed polygon of the target non-grass area, and marking the circumscribed polygon as the non-passable area. That is, obtaining the convex hull of the fence area in the image.
[0050] Furthermore, after determining the circumscribed polygon of the target non-grass area, the method may further include: enlarging the circumscribed polygon at the same distance to obtain the enlarged second area, and marking the second area as a non-passable area. The equidistant magnification coefficient is preferably 25 pixels. The precision of image magnification is less than or equal to the precision of image expansion. The equidistant magnification coefficient is related to the suspended height of the fence. When the suspended bottom of the fence is higher from the ground, the corresponding magnification coefficient is larger, and the enlarged fence area is closer to the real boundary, that is, closer to the projection of the fence on the ground. The equidistant magnification coefficient can be a parameter pre-set in the lawn mower. After obtaining the circumscribed polygon, the circumscribed polygon can be directly enlarged by the parameter to obtain the non-passable area. The equidistant magnification coefficient can also be set by the user according to the suspended height of his own fence before the lawn mower starts working, and the circumscribed polygon is enlarged according to the set parameter to obtain the non-passable area.
[0051] Due to errors and other reasons in the image processing process, a target non-grass area may be divided into multiple areas after segmentation. If multiple target non-grass areas are directly processed at this time, the result is obviously inaccurate. Therefore, in an embodiment of the present application, before determining the circumscribed polygon of the target non-grass area, the images that may come from the same target non-grass area may be connected (merged) first, and then the above-mentioned expansion or convex hull search and enlargement operations may be performed after the connection (merging) to improve the image segmentation accuracy. Of course, only the merging process may be performed to treat the area obtained after the merging process as a non-passable area, or the above-mentioned expansion or convex hull search and enlargement operations may be performed on the connected (merged) area, and the area obtained after various operations may be treated as a non-passable area. As long as the obtained area has high accuracy, this application does not limit the order of steps and the specific operations to be used.
[0052] Specifically, Figure 3 As shown in the figure, when at least two target non-grass areas are identified (a first target non-grass area 22 and a second target non-grass area 23), the target non-grass area 22 with the largest area and the target non-grass area 23 that meets the requirements are merged, and the merged area is marked as a non-passable area 24. There are pixels located in the lower half of the image and in the lower right corner of the image in the target non-grass area that meets the requirements.
[0053] In one embodiment of the present application, the central axis of the image in the horizontal direction and the central axis of the lawn mower are obtained respectively, and the distance between the pixel point located in the lower half of the image and the right side of the lawn mower is smaller than the distance between the pixel point located in the upper half and the right side of the lawn mower. The central axis of the image in the vertical direction is obtained, and the distance between the pixel point located in the lower right corner of the image and the right side of the lawn mower is smaller than the distance between the pixel point located in the lower left corner and the right side of the lawn mower. In another embodiment of the present application, the target non-grass area that meets the requirements has a pixel point with a horizontal coordinate of 511 and a vertical coordinate greater than or equal to 256 in the image coordinate system.
[0054] In the embodiment of the present application, after identifying multiple target non-grass areas, first obtain the area with the largest area as the connected (merged) main part, then identify the pixel points located in the lower half of the image and in the lower right corner of the image, and connect (merge) them with the main part. The circumscribed polygon is enlarged by the same distance to obtain the enlarged second area, and the second image is marked as a non-passable area.
[0055] The above two methods of expanding the target non-grass area can be used in the process of the lawn mower driving along the edge (including returning to charge) and the process of random driving. When a target non-grass area that meets the requirements such as a suspended fence is identified during the process of driving along the edge or randomly driving, it is processed accordingly to obtain a boundary that is closer to the actual boundary.
[0056] In an embodiment of the present application, the target non-grass area is expanded so that the boundary of the area is close to the actual physical boundary, so that the lawn mower is less likely to go out of bounds, the camera is stuck, and other unsafe phenomena occur when driving according to the actual physical boundary.
[0057] S404: Controlling the driving and / or operation of the self-moving device according to the calibration result.
[0058] In this embodiment, in the edge mode and the random mode, the lawn mower travels and / or works in the traversable area according to the calibration result.
[0059] An automatic working method is proposed in an embodiment of the present application. In this method, the segmented areas such as the fence are expanded so that the boundaries of the expanded fence area match the physical boundaries determined by the fence in the actual scene. When the lawn mower travels according to the image processed by the automatic working method, it will travel in the working area determined by the actual physical boundaries, thereby avoiding safety problems such as the lawn mower driving out of the boundary.
[0060] An embodiment of the present invention further provides a self-moving device, which may include: a collection module, which may be used to collect an image in front of the self-moving device;
[0061] An image segmentation module, which can be used to segment the image to obtain a segmentation result, wherein the segmentation result includes a target non-grass area;
[0062] An image expansion module may be used to expand the target non-grass area and mark the area determined after the expansion as a non-passable area, wherein the non-passable area is an area where the self-moving device cannot travel and / or work;
[0063] The mowing driving module can be used to control the driving and / or operation of the self-moving device according to the calibration results.
[0064] In one embodiment, the target non-grass area includes at least one of the following: a fence area, a fence area, and there is a gap in the vertical direction between the fence or a part of the bottom of the fence and the boundary.
[0065] In one embodiment, the image expansion module may include: performing image expansion on the target non-grass area according to a preset rule to obtain an expanded first area, and marking the first area as a non-passable area.
[0066] In one embodiment, the preset rule includes: expanding the target non-grass area up and down and / or left and right.
[0067] In one embodiment, the preset rule includes: expanding the target non-grass area up and down and expanding the target non-grass area left and right, wherein the pixel points of the up and down expansion and the pixel points of the left and right expansion are different.
[0068] In one embodiment, the image expansion module may include: determining a circumscribed polygon of the target non-grass area, and marking the circumscribed polygon as the non-passable area.
[0069] In one embodiment, before determining the circumscribed polygon of the target non-grass area, the image expansion module may further include:
[0070] When at least two target non-grass areas are identified, the target non-grass area with the largest area and the target non-grass area that meets the requirements are merged, and the merged area is marked as the non-passable area. Among the target non-grass areas that meet the requirements, there are pixels located in the lower half of the image and in the lower right corner of the image.
[0071] In one embodiment, the target non-grass area that meets the requirements has a pixel point whose abscissa is 511 and whose ordinate is greater than or equal to 256 in the image coordinate system.
[0072] In one embodiment, after determining the circumscribed polygon of the target non-grass area, the image expansion module may further include:
[0073] The circumscribed polygon is enlarged by the same distance to obtain an enlarged second area, and the second area is marked as a non-passable area, and the distance is related to the suspended height of the target non-grass area.
[0074] In one embodiment, the distance is set by a user.
[0075] In addition to the above-mentioned methods and devices, an embodiment of the present application may also be a computer program product, which includes computer program instructions, which, when executed by a processor, enable the processor to execute the steps of the method for determining the direction of a mobile device according to various embodiments of the present invention described in the above-mentioned "Exemplary Direction Determination Method" section of this specification.
[0076] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, a plurality of steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.
[0077] A person skilled in the art may understand that all or part of the steps in the method for implementing the above-mentioned embodiment may be completed by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, which, when executed, includes one or a combination of the steps of the method embodiment.
[0078] In addition, each functional unit in each embodiment of the present invention may be integrated into a processing module, or each unit may exist physically separately, or two or more units may be integrated into one module. The above-mentioned integrated module may be implemented in the form of hardware or in the form of a software functional module. If the integrated module is implemented in the form of a software functional module and sold or used as an independent product, it may also be stored in a computer-readable storage medium.
[0079] The storage medium mentioned above can be a read-only memory, a magnetic disk or an optical disk, etc.
[0080] Although only a few embodiments of the present invention are described and illustrated in this specification, those skilled in the art should be able to easily foresee other means or structures for performing the functions and / or obtaining the structures described herein, and each such change or modification is considered to be within the scope of the present invention.
Claims
1. An automatic working method, characterized in that: The method is applied to a self-moving device that travels and / or works in a work area defined by boundaries, and the method comprises: Capturing the image in front of the mobile device; Segmenting the image to obtain a segmentation result, wherein the segmentation result includes a target non-grass area; Expanding the target non-grass area, and marking the area determined after the expansion as a non-passable area, wherein the non-passable area is an area where the self-moving device cannot travel and / or work; The driving and / or operation of the self-moving device is controlled according to the calibration result.
2. The automatic working method according to claim 1, characterized in that: The target non-grass area includes at least one of the following: a fence area, a fence area, and there is a gap in the vertical direction between the fence or a part of the bottom of the fence and the boundary.
3. The automatic working method according to claim 1, characterized in that: The target non-grass area includes: The target non-grass area is image-expanded according to a preset rule to obtain an expanded first area, and the first area is marked as a non-passable area.
4. The automatic working method according to claim 3, characterized in that: The preset rules include: vertically expanding and / or horizontally expanding the target non-grass area.
5. The automatic working method according to claim 4, characterized in that: The pixels of the vertical expansion and the horizontal expansion are different.
6. The automatic working method according to claim 1, characterized in that: The target non-grass area includes: A circumscribed polygon of the target non-grass area is determined, and the circumscribed polygon is marked as the non-passable area.
7. The automatic working method according to claim 6, characterized in that: Before determining the circumscribed polygon of the target non-grass area, the method further includes: When at least two target non-grass areas are identified, the target non-grass area with the largest connected area and the target non-grass area that meets the requirements are merged and marked as the non-passable area. Among the target non-grass areas that meet the requirements, there are pixels located in the lower half of the image and in the lower right corner of the image.
8. The automatic working method according to claim 6, characterized in that: After determining the circumscribed polygon of the target non-grass area, the method further includes: The circumscribed polygon is enlarged by the same distance to obtain an enlarged second area, and the second area is marked as a non-passable area, and the distance is related to the suspended height of the target non-grass area.
9. The automatic working method according to claim 8, characterized in that: The distance is set by the user.
10. A self-propelled device, characterized in that: The self-mobile device comprises: A visual sensor, the visual sensor is used to collect the image in front of the mobile device; A controller is electrically connected to the visual sensor, and the controller segments the image to obtain a segmentation result, wherein the segmentation result includes a target non-grass area; the target non-grass area is expanded, and the area determined after the expansion is calibrated as a non-passable area, wherein the non-passable area is an area where the self-moving device cannot travel and / or work, and the travel and / or work of the self-moving device is controlled according to the calibration result.
