Grid map-based connected domain corridor optimal full coverage path planning method and system

Abstract

The application discloses a grid map connected domain corridor optimal full coverage path planning method and system, comprising the following steps: first, a two-dimensional grid map of a lawn environment is acquired, and image processing is performed on the two-dimensional grid map to generate a fine-granularity binary grayscale grid map; second, based on an adaptive resolution adjustment rule, the two directions of the coordinate system of the binary grayscale grid map are compressed and adjusted respectively to generate a compressed two-dimensional array map; third, based on a rectangular connected domain division rule, the compressed two-dimensional array map is divided and updated to dynamically generate a rectangular connected domain set; and fourth, for each rectangular connected domain in the rectangular connected domain set, a ploughing type full coverage path planning algorithm is adopted to generate a travel path in a single rectangular connected domain, and for the rectangular connected domains in the rectangular connected domain set, an A* heuristic algorithm is used to search for paths between the rectangular connected domains and generate a safe corridor between the rectangular connected domains.

Description

Technical Field

[0001] This invention relates to the field of robot path planning technology, and in particular to a method and system for connected domain partitioning and full-coverage path planning based on grid maps. Background Technology

[0002] With the development of robotics technology, especially its increasingly widespread application in industry, military, and service sectors, higher demands are being placed on robots' path planning capabilities. In complex environments, how to efficiently and safely plan a fully covered path has become an important research direction in the field of robotics. Traditional path planning methods often rely on simple map models and fixed algorithms, making it difficult to adapt to changing environments and complex task requirements.

[0003] Therefore, there is an urgent need for a method and system for optimal full-coverage path planning based on grid map connected domain corridors, which can intelligently identify environmental features, divide connected domains, and generate safe corridors, and is of great significance for improving the autonomy and adaptability of robots. Summary of the Invention

[0004] This invention addresses the problems and shortcomings of existing technologies by providing a novel method and system for optimal full-coverage path planning of connected domain corridors based on grid maps.

[0005] The present invention solves the above-mentioned technical problems through the following technical solution:

[0006] This invention provides a method for optimal full-coverage path planning based on connected domain corridors in a raster map, characterized by including:

[0007] S1. Two-dimensional raster map image preprocessing: Obtain a two-dimensional raster map of the lawn environment, and perform image processing on the two-dimensional raster map to generate a fine-grained binary grayscale raster map.

[0008] S2. Two-dimensional raster map adaptive resolution adjustment: Based on the adaptive resolution adjustment rules, the coordinate system of the binary grayscale raster map is compressed and adjusted in two directions respectively to generate a compressed two-dimensional array map;

[0009] S3. Rectangular Connected Component Partitioning: Based on the rectangular connected component partitioning rules, the compressed two-dimensional array map is partitioned and updated to dynamically generate a set of rectangular connected components;

[0010] S4. Optimal Full Coverage Path Planning: For each rectangular connected region in the set of rectangular connected regions, the ox-plowing full coverage path planning algorithm is used to generate the travel path in a single rectangular connected region. For the rectangular connected regions in the set of rectangular connected regions, the heuristic algorithm of A* is used to search for the path between the rectangular connected regions and generate the safe corridor between the rectangular connected regions, thereby constructing the optimal full coverage planning path.

[0011] This invention also provides a grid map-based optimal full-coverage path planning system for connected domain corridors, characterized by comprising:

[0012] A map preprocessing module: acquires a two-dimensional raster map of the lawn environment, and performs image processing on the two-dimensional raster map to generate a fine-grained binary grayscale raster map.

[0013] A map scaling adjustment module: Based on adaptive resolution adjustment rules, the coordinate system of the binary grayscale raster map is compressed and adjusted in two directions to generate a compressed two-dimensional array map;

[0014] A rectangular connected component partitioning module: Based on the rectangular connected component partitioning rules, the compressed two-dimensional array map is analyzed by traversing column by column, and the set of rectangular connected components is dynamically generated and updated.

[0015] An optimal full-coverage path planning module: For each rectangular connected region in the set of rectangular connected regions, the ox-plowing full-coverage path planning algorithm is used to generate the travel path in a single rectangular connected region. For the rectangular connected regions in the set of rectangular connected regions, the heuristic algorithm of A* is used to search for the path between the rectangular connected regions and generate the safe corridor between the rectangular connected regions, thereby constructing the optimal full-coverage planning path.

[0016] The positive and progressive effects of this invention are as follows:

[0017] The present invention provides a method and system for optimal full-coverage path planning based on grid map connected domain corridors. It can refine the processing of two-dimensional grid maps and intelligently divide the two-dimensional grid maps into rectangular connected domains. The rectangular connected domains are fully covered by the ox-plowing full-coverage path planning algorithm, and safe corridors between the rectangular connected domains are generated using the A* heuristic algorithm, so as to improve the motion planning and execution efficiency of robots in complex environments. Attached Figure Description

[0018] Figure 1 This is a flowchart illustrating the preferred embodiment of the optimal full-coverage path planning method for connected domain corridors based on a grid map according to the present invention.

[0019] Figure 2This is a diagram illustrating the effect of dividing and updating rectangular connected components based on a compressed two-dimensional array map, representing a preferred embodiment of the present invention.

[0020] Figure 3 This is a schematic diagram of the process for preprocessing two-dimensional raster map images according to a preferred embodiment of the present invention.

[0021] Figure 4 This is a diagram illustrating the optimal full-coverage path planning of a preferred embodiment of the present invention.

[0022] Figure 5 This is a schematic diagram of the structure of a grid map-based connected domain corridor optimal full-coverage path planning system according to a preferred embodiment of the present invention. Detailed Implementation

[0023] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0024] like Figure 1 and Figure 2 As shown in this embodiment, the optimal full-coverage path planning method for corridors in connected domains of a raster map includes:

[0025] S1. Two-dimensional raster map image preprocessing: Obtain a two-dimensional raster map of the lawn environment, and perform image processing on the two-dimensional raster map to generate a fine-grained binary grayscale raster map.

[0026] In this step, the two-dimensional raster map of the lawn environment is a two-dimensional planar composition compressed from a three-dimensional image scanned by a LiDAR sensor within a specified height range. By performing edge-bending processing on the obstacle boundaries and image edge dilation processing on the two-dimensional raster map image, a binarized grayscale raster map is obtained.

[0027] In this step, by refining and recognizing the input raster map, we can better understand and represent environmental features. The innovation lies in achieving accurate identification and generation of map granularity, which provides a solid foundation for subsequent intelligent decision-making based on environmental features.

[0028] S2. Adaptive resolution adjustment of 2D raster map: Based on the adaptive resolution adjustment rules, the coordinate system of the binary grayscale raster map is compressed and adjusted in both directions to generate a compressed 2D array map.

[0029] In this step, the adaptive resolution adjustment rule calculates the horizontal and vertical scaling factors for the two directions of the coordinate system of the binary grayscale raster map, and then compresses and adjusts the coordinate system in both directions according to the horizontal and vertical scaling factors to generate a compressed two-dimensional array map. Wherein:

[0030] The calculation rule for the horizontal scaling factor is as follows:

[0031] Calculate the first projected length of each obstacle on the horizontal coordinate of the two-dimensional grid map;

[0032] Obtain the minimum first projection length;

[0033] When the minimum first projection length is less than the preset scaling threshold, the horizontal scaling factor is the minimum first projection length divided by the initial resolution.

[0034] When the minimum first projection length is greater than or equal to the preset scaling threshold, the horizontal scaling factor is the preset scaling threshold divided by the initial resolution, where the initial resolution is the resolution corresponding to the two-dimensional raster map obtained in S1.

[0035] The calculation rule for the vertical scaling factor is as follows:

[0036] Calculate the second projection length of each obstacle on the vertical coordinate of the two-dimensional grid map;

[0037] Obtain the minimum second projection length;

[0038] When the minimum second projection length is less than the preset scaling threshold, the vertical scaling factor is the minimum second projection length divided by the initial resolution;

[0039] When the minimum second projection length is greater than or equal to a preset scaling threshold, the vertical scaling factor is the preset scaling threshold divided by the initial resolution, where the initial resolution is the resolution corresponding to the two-dimensional raster map obtained in S1. In this embodiment, the preset scaling threshold is 5cm.

[0040] After map image processing, considering the typically large size of obstacles in lawn environments, using the initial resolution (e.g., an initial image pixel to map ratio of 1 pixel: 5 millimeters) would result in an excessively large map size and excessive memory consumption. Since the overall map is a two-dimensional plane, this invention calculates the number of pixels projected onto the corresponding axes (horizontal and vertical) to determine the minimum unit pixel size on each axis. For example, on the horizontal axis, the area occupied by each obstacle's projection is calculated. When the minimum length is small (less than 5cm), this minimum length is divided by the initial resolution as the scaling factor for the horizontal axis. When the minimum length is large (greater than 5cm), a preset resolution interval (5cm) is used, with 5cm as the dividing line, divided by the initial resolution as the scaling factor. The horizontal axis of the two-dimensional map is compressed according to the determined scaling factor. This significantly reduces storage space, adapts to a wide range of lawn care environments, and does not significantly affect the overall obstacle resolution. In the implemented instances and scenarios, the binarized grayscale raster map achieved 10x compression on both the horizontal and vertical axes of the coordinate system, reducing the overall storage footprint to 1% of the original. This also saved time for subsequent map searches.

[0041] S3. Rectangular Connected Component Partition: Based on the rectangular connected component partitioning rules, the compressed two-dimensional array map is partitioned and updated to dynamically generate a set of rectangular connected components.

[0042] In this step, the rules for dividing the rectangular connected regions can be found in Examples 1 and 2.

[0043] Example 1: The compressed two-dimensional array map is traversed column by column. The column index of each column is read in turn. In each column, each row of pixels is traversed from top to bottom. Rectangular connected components are divided based on the pixels. After the column-by-column traversal and division, the adjacent rectangular connected components are merged and updated to dynamically generate a set of rectangular connected components.

[0044] Specifically, the compressed two-dimensional array map is traversed column by column, and the column index of each column is read in turn. In each column, each row of pixels is traversed from top to bottom. In the compressed two-dimensional array map, 0 represents blank land (i.e., reachable land) and 1 represents obstacles (i.e., unreachable land).

[0045] Analyze the current pixel in the current column:

[0046] If the current pixel is the first pixel in the first row and is a reachable location 0, then a new rectangular connected region is constructed based on the current pixel, the first row is marked as the starting row index of the newly constructed rectangular connected region, and the next pixel in the current column is traversed. If the current pixel is the first pixel in the first row and is an inaccessible location 1, then the first row cannot be used as the starting row index, and the next pixel in the current column is traversed directly.

[0047] If the current pixel is not the first pixel in the row and both the previous pixel and the current pixel are reachable grounds (0), then the range of the rectangular connected region built based on the previous pixel is extended to the current pixel. If the current pixel is not the last pixel in the row, then the traversal continues to the next pixel in the current column. If the current pixel is the last pixel in the row but not the last column, then the current rectangular connected region ends, the current last pixel is marked as the endpoint row index, and the traversal continues to the next column. If the current pixel is the last pixel in the row and the last column, then the traversal ends. If the current pixel is not the first pixel in the row and is reachable ground (0), and the previous pixel is an inaccessible ground (1), then a new rectangular connected region is built based on the current pixel. The current row is marked as the starting row index of the newly built rectangular connected region. If the current pixel is not the last pixel in the row, then the traversal continues to the next pixel in the current column. If the current pixel is the last pixel in the row but not the last column, then the traversal continues to the next column (the new connected region built based on the current pixel only has a starting row index and no endpoint row index, so it does not constitute a true rectangular connected region). If the current pixel is the last pixel in the row and the last column, then the traversal ends.

[0048] If the current pixel is not the first pixel in the row and is an unreachable location 1, and the previous pixel is a reachable location 0, then the rectangular connected component constructed based on the previous pixel ends, the row index of the previous pixel is marked as the ending row index, and if the current pixel is not the last pixel in the row, then continue traversing the next pixel in the current column; if the current pixel is the last pixel in the row but not the last pixel in the column, then continue traversing the next column; if the current pixel is the last pixel in the row and the last pixel in the column, then the traversal ends. If the current pixel is not the first pixel in the row and both the previous pixel and the previous pixel are unreachable locations 1, then if the current pixel is not the last pixel in the row, then continue traversing the next pixel in the current column; if the current pixel is the last pixel in the row but not the last pixel in the column, then continue traversing the next column; if the current pixel is the last pixel in the row and the last pixel in the column, then the traversal ends.

[0049] After traversing each column, obtain the rectangular connected region to be updated for each column, which has a starting row index and an ending row index, and the starting row index and the ending row index are different.

[0050] For the rectangular connected components to be updated, compare the rectangular connected components of the current column with the rectangular connected components of the previous column to determine whether the connected component update conditions are met. If they are, merge and update the rectangular connected components; otherwise, do not update them. Finally, obtain the set of rectangular connected components, with the current column being the second column to the last column in sequence.

[0051] The condition for updating the connected component is that when the starting row index and ending row index of the rectangular connected component in the current column are exactly the same as the starting row index and ending row index of the rectangular connected component in the previous column, the rectangular connected component in the current column and the rectangular connected component in the previous column are merged into one rectangular connected component.

[0052] Example 2: The compressed two-dimensional array map is traversed column by column. The column index of each column is read in turn. In each column, each row of pixels is traversed from top to bottom. Rectangular connected components are divided based on the pixels. During the column-by-column traversal and division process, adjacent rectangular connected components are merged and updated to dynamically generate a set of rectangular connected components.

[0053] Specifically, the compressed two-dimensional array map is traversed column by column, and the column index of each column is read in turn. In each column, each row of pixels is traversed from top to bottom. In the compressed two-dimensional array map, 0 represents blank land (i.e., reachable land) and 1 represents obstacles (i.e., unreachable land).

[0054] Analyze the current pixel in the current column:

[0055] For the current column being the first column: if the current pixel is the first row pixel and is a reachable location 0, then a new rectangular connected region is constructed based on the current pixel, the current first row is marked as the starting row index of the newly constructed rectangular connected region, and the next pixel in the current column is traversed; if the current pixel is the first row pixel and is an inaccessible location 1, then the first row cannot be used as the starting row index, and the next pixel in the current column is traversed directly.

[0056] If the current pixel is not the first pixel in the row and both the previous pixel and the current pixel are reachable from the ground 0, then the range of the rectangular connected region built based on the previous pixel is extended to the current pixel. If the current pixel is not the last pixel in the row, then the next pixel in the current column is traversed. If the current pixel is the last pixel in the row, then the current rectangular connected region ends, the current last pixel is marked as the endpoint row index, and the next column is traversed. If the current pixel is not the first pixel in the row and is reachable from the ground 0, and the previous pixel is inaccessible from the ground 1, then a new rectangular connected region is built based on the current pixel, the current row is marked as the starting row index of the newly built rectangular connected region, if the current pixel is not the last pixel in the row, then the next pixel in the current column is traversed. If the current pixel is the last pixel in the row, the next column is traversed. (The new connected region built based on the current pixel only has a starting row index and no endpoint row index, so it does not constitute a true rectangular connected region.)

[0057] If the current pixel is not the first pixel in the row and is an unreachable location 1, and the previous pixel is a reachable location 0, then the rectangular connected component constructed based on the previous pixel ends, and the row index of the previous pixel is marked as the ending row index. If the current pixel is not the last pixel in the row, then continue traversing the next pixel in the current column; if the current pixel is the last pixel in the row, then continue traversing the next column. If the current pixel is not the first pixel in the row and both the previous pixel and the previous pixel are unreachable locations 1, then continue traversing the next pixel in the current column if the current pixel is not the last pixel in the row; if the current pixel is the last pixel in the row, then continue traversing the next column.

[0058] For cases where the current column is not the first column: if the current pixel is the first row pixel and is reachable (0), then a new rectangular connected region is constructed based on the current pixel, the current first row is marked as the starting row index of the newly constructed rectangular connected region, and the next pixel in the current column is traversed; if the current pixel is the first row pixel and is unreachable (1), then the first row cannot be used as the starting row index, and the next pixel in the current column is traversed directly.

[0059] If the current pixel is not the first pixel in the row and both the current pixel and the previous pixel are reachable from the ground 0, then the range of the rectangular connected region built based on the previous pixel is extended to the current pixel. If the current pixel is not the last pixel in the row, the process continues to traverse the next pixel in the current column. If the current pixel is the last pixel in the row, the current rectangular connected region ends, and the current last pixel is marked as the endpoint row index. The process analyzes whether the current column and the previous column meet the connected region update conditions. If they do, the update is performed; otherwise, no update is performed, and the process continues to traverse the next column. If the current pixel is not the first pixel in the row and is reachable from the ground 0, and the previous pixel is inaccessible from the ground 1, then a new rectangular connected region is built based on the current pixel. The current row is marked as the starting row index of the newly built rectangular connected region. If the current pixel is not the last pixel in the row, the process continues to traverse the next pixel in the current column. If the current pixel is the last pixel in the row (the new connected region built based on the current pixel only has a starting row index and no endpoint row index, so it does not constitute a true rectangular connected region), the process analyzes whether the current column and the previous column meet the connected region update conditions. If they do, the update is performed; otherwise, no update is performed, and the process continues to traverse the next column.

[0060] If the current pixel is not the first pixel in the row and is an unreachable location 1, and the previous pixel is a reachable location 0, then the rectangular connected component constructed based on the previous pixel ends, and the row index of the previous pixel is marked as the ending row index. If the current pixel is not the last pixel in the row, then continue traversing the next pixel in the current column. If the current pixel is the last pixel in the row, then analyze whether the current column and the previous column satisfy the connected component update condition. If yes, update; otherwise, do not update, and continue traversing the next column. If the current pixel is not the first pixel in the row and is an unreachable location 1 in the row, then continue traversing the next pixel in the current column. If the current pixel is the last pixel in the row, then analyze whether the current column and the previous column satisfy the connected component update condition. If yes, update; otherwise, do not update, and continue traversing the next column.

[0061] After traversing each column, we obtain a set of rectangular connected regions that have been partitioned and updated, with a starting row index and an ending row index that are different from each other.

[0062] The condition for updating the connected component is that when the starting row index and ending row index of the rectangular connected component in the current column are exactly the same as the starting row index and ending row index of the rectangular connected component in the previous column, the rectangular connected component in the current column and the rectangular connected component in the previous column are merged into one rectangular connected component.

[0063] In this step, the connected domains in the environment can be intelligently analyzed and divided. By using connectivity information to divide the environment, a more accurate and efficient environmental coverage scheme can be provided for the system, thereby improving the adaptability of the intelligent system in complex environments.

[0064] S4, Optimal Full Coverage Path Planning (see...) Figure 4 For each rectangular connected region in the set of rectangular connected regions, the ox-plowing full-coverage path planning algorithm is used to generate the travel path in a single rectangular connected region. For the rectangular connected regions in the set of rectangular connected regions, the heuristic algorithm of A* is used to search for the path between the rectangular connected regions and generate the safe corridor between the rectangular connected regions, thereby constructing the optimal full-coverage planning path.

[0065] In this step, a rudimentary path planning algorithm covering all accessible areas within rectangular connected regions is employed to traverse and cover all reachable regions of each rectangular connected region sequentially. Simultaneously, the A* heuristic algorithm is used to plan the shortest connection paths between different rectangular connected regions, enabling efficient generation of complete paths covering all passable areas from the binary grayscale raster map after the rectangular connected regions are divided.

[0066] Ox-plowing Full-Coverage Path Planning Algorithm within Rectangular Connected Domains: Within each identified rectangular connected domain, the ox-plowing full-coverage path planning algorithm is used to generate a full-coverage path within that rectangular connected domain. This algorithm systematically covers every part of each rectangular connected domain by simulating an ox plowing a field, ensuring the comprehensiveness and efficiency of path planning.

[0067] A heuristic algorithm for A* between rectangular connected regions: After completing the full-coverage path planning within each rectangular connected region, the A* heuristic algorithm is used to generate safe corridors between the rectangular connected regions, achieving global full-coverage path planning. The A* heuristic algorithm is used to find the shortest path planning from the starting point to the ending point in a binary grayscale raster map.

[0068] In this step, safe corridors between connected domains can be intelligently planned and generated to ensure the safe movement and navigation of robots or intelligent systems in the environment. Combined with the path planning capabilities of the A* algorithm, safe corridors between connected domains are effectively generated, improving the motion planning and execution efficiency of intelligent systems in complex environments.

[0069] like Figure 3 As shown, the two-dimensional raster map image preprocessing provided in this embodiment includes:

[0070] S11. Perform grayscale processing on the two-dimensional raster map, including removing outliers and smoothing obstacle edges.

[0071] In this step, the 2D raster map image contains a large number of noise points and tortuous obstacle edges, causing redundant cutting in the subsequent connected component cutting task. To achieve better connected component cutting results and finer-grained map recognition, this invention utilizes image edge-smoothing techniques from computational visual image processing to remove outliers and smooth obstacle edges on the 2D raster map image. The purpose is to:

[0072] 1. Sharpen the edge of obstacles to overcome the problem of large tortuous fluctuations caused by outliers at the edge of obstacles.

[0073] 2. Eliminate the influence of outliers. Due to factors such as sensor data acquisition, there will be many outliers in two-dimensional grid maps. Image edge processing can filter out discontinuous outliers, thereby improving the overall accuracy of two-dimensional grid maps.

[0074] S12. Expand the edges of obstacles in the processed map according to the set expansion radius.

[0075] In this step, after the 2D raster map is processed into a grayscale image, image edge dilation is performed. To avoid the loss of information due to the edge-dilation of obstacle boundaries, image edge dilation technology is further employed. By setting safe collision parameters, the edges of obstacles are simultaneously dilated by a relative distance to obtain the processed map. Obstacle edge dilation ensures safe obstacle avoidance for the robot during actual execution.

[0076] S13. Generate a fine-grained binary grayscale raster map.

[0077] like Figure 5 As shown, the raster map connected domain corridor optimal full coverage path planning system 500 provided in this embodiment includes:

[0078] A map preprocessing module 501: acquires a two-dimensional raster map of the lawn environment, performs image processing on the two-dimensional raster map to generate a fine-grained binary grayscale raster map.

[0079] A map scaling adjustment module 502: Based on the adaptive resolution adjustment rules, the coordinate system of the binary grayscale raster map is compressed and adjusted in two directions to generate a compressed two-dimensional array map.

[0080] Module 503 for dividing rectangular connected components: Based on the rules for dividing rectangular connected components, it uses a column-by-column traversal method to analyze the compressed two-dimensional array map, and dynamically generates and updates the set of rectangular connected components.

[0081] Module 504: For each rectangular connected region in the set of rectangular connected regions, the ox-plowing full-coverage path planning algorithm is used to generate the travel path in a single rectangular connected region. For the rectangular connected regions in the set of rectangular connected regions, the heuristic algorithm of A* is used to search for the path between the rectangular connected regions and generate the safe corridor between the rectangular connected regions, thereby constructing the optimal full-coverage planning path.

[0082] This invention relates to the fine-grained processing and recognition generation of raster maps, enabling detailed analysis and recognition of raster maps in the environment to better understand and represent environmental features. The innovation lies in achieving accurate identification and generation of map granularity, providing a solid foundation for subsequent intelligent decision-making based on environmental features. This invention also relates to methods for connected component analysis and intelligent partitioning of covered environments. It can intelligently analyze and partition connected components in the environment to better understand the layout and structure. The key is utilizing connectivity information to partition the environment, thereby providing the system with a more accurate and efficient environmental coverage scheme and improving the adaptability of intelligent systems in complex environments. Finally, this invention relates to generating safe corridors between connected components using the A* algorithm. It can intelligently plan and generate safe corridors between connected components to ensure the safe movement and navigation of robots or intelligent systems in the environment. Combined with the path planning capabilities of the A* algorithm, it effectively generates safe corridors between connected components, improving the motion planning and execution efficiency of intelligent systems in complex environments. In summary, this invention provides a comprehensive environmental perception and path planning solution, providing effective support for the movement, decision-making, and navigation of intelligent systems in complex environments, and has broad application prospects and technical value.

[0083] While specific embodiments of the present invention have been described above, those skilled in the art should understand that these are merely illustrative examples, and the scope of protection of the present invention is defined by the appended claims. Those skilled in the art can make various changes or modifications to these embodiments without departing from the principles and essence of the present invention, but all such changes and modifications fall within the scope of protection of the present invention.

Claims

1. A method for optimal full-coverage path planning in connected domains of a raster map, characterized in that, It includes: S1. Two-dimensional raster map image preprocessing: Obtain a two-dimensional raster map of the lawn environment, and perform image processing on the two-dimensional raster map to generate a fine-grained binary grayscale raster map. S2. Two-dimensional raster map adaptive resolution adjustment: Based on the adaptive resolution adjustment rules, the coordinate system of the binary grayscale raster map is compressed and adjusted in two directions respectively to generate a compressed two-dimensional array map; S3. Rectangular Connected Component Partitioning: Based on the rectangular connected component partitioning rules, the compressed two-dimensional array map is partitioned and updated to dynamically generate a set of rectangular connected components; In S3, the rule for partitioning the rectangular connected region is as follows: The compressed two-dimensional array map is traversed column by column, and the column index of each column is read in turn. In each column, each row of pixels is traversed from top to bottom, and rectangular connected components are divided based on the pixels. After the column-by-column traversal and division, the adjacent rectangular connected components are merged and updated to dynamically generate a set of rectangular connected components. Alternatively, during the column-by-column traversal and division process, the adjacent rectangular connected components are merged and updated to dynamically generate a set of rectangular connected components. S4. Optimal Full Coverage Path Planning: For each rectangular connected region in the set of rectangular connected regions, the ox-plowing full coverage path planning algorithm is used to generate the travel path in a single rectangular connected region. For the rectangular connected regions in the set of rectangular connected regions, the heuristic algorithm of A* is used to search for the path between the rectangular connected regions and generate the safe corridor between the rectangular connected regions, thereby constructing the optimal full coverage planning path.

2. The optimal full-coverage path planning method for corridors based on grid map connected domains as described in claim 1, characterized in that, In S2, the adaptive resolution adjustment rule is to calculate the horizontal scaling factor and the vertical scaling factor for the two directions of the coordinate system of the binary grayscale raster map, and to compress and adjust the two directions of the coordinate system according to the horizontal scaling factor and the vertical scaling factor to generate a compressed two-dimensional array map.

3. The optimal full-coverage path planning method for corridors based on grid map connected domains as described in claim 2, characterized in that, The calculation rule for the horizontal scaling factor is as follows: Calculate the first projected length of each obstacle on the horizontal coordinate of the two-dimensional grid map; Obtain the minimum first projection length; When the minimum first projection length is less than the preset scaling threshold, the horizontal scaling factor is the minimum first projection length divided by the initial resolution; When the minimum first projection length is greater than or equal to the preset scaling threshold, the horizontal scaling factor is the preset scaling threshold divided by the initial resolution, where the initial resolution is the resolution corresponding to the two-dimensional raster map obtained in S1.

4. The optimal full-coverage path planning method for corridors based on grid map connected domains as described in claim 2, characterized in that, The calculation rule for the vertical scaling factor is as follows: Calculate the second projection length of each obstacle on the vertical coordinate of the two-dimensional grid map; Obtain the minimum second projection length; When the minimum second projection length is less than the preset scaling threshold, the vertical scaling factor is the minimum second projection length divided by the initial resolution; When the minimum second projection length is greater than or equal to the preset scaling threshold, the vertical scaling factor is the preset scaling threshold divided by the initial resolution, where the initial resolution is the resolution corresponding to the two-dimensional raster map obtained in S1.

5. The optimal full-coverage path planning method for corridors based on grid map connected domains as described in claim 1, characterized in that, In S3, the rule for partitioning the rectangular connected region is as follows: The compressed two-dimensional array map is traversed column by column, and the column index of each column is read in turn. In each column, each row of pixels is traversed from top to bottom. In the compressed two-dimensional array map, 0 represents empty land that can be reached, and 1 represents obstacles that are inaccessible. Analyze the current pixel in the current column: If the current pixel is the first pixel in the first row and is a reachable location 0, then a new rectangular connected region is constructed based on the current pixel, the first row is marked as the starting row index of the newly constructed rectangular connected region, and the next pixel in the current column is traversed. If the current pixel is the first pixel in the first row and is an inaccessible location 1, then the first row cannot be used as the starting row index, and the next pixel in the current column is traversed directly. If the current pixel is not the first pixel in the row and both the previous pixel and the current pixel are reachable from the ground 0, then the range of the rectangular connected region built based on the previous pixel is extended to the current pixel. If the current pixel is not the last pixel in the row, then the traversal continues to the next pixel in the current column. If the current pixel is the last pixel in the row but not the last pixel in the column, then the current rectangular connected region ends, the current last pixel is marked as the index of the end row, and the traversal continues to the next column. If the current pixel is the last pixel in the row and the last pixel in the column, then the traversal ends. If the current pixel is not the first pixel in the row and is reachable from the ground 0, and the previous pixel is inaccessible from the ground 1, then a new rectangular connected region is built based on the current pixel, the current row is marked as the index of the starting row of the newly built rectangular connected region, if the current pixel is not the last pixel in the row, then the traversal continues to the next pixel in the current column. If the current pixel is the last pixel in the row but not the last pixel in the column, then the traversal continues to the next column. If the current pixel is the last pixel in the row and the last pixel in the column, then the traversal ends. If the current pixel is not the first pixel and is an unreachable location 1, and the previous pixel is a reachable location 0, then the rectangular connected component constructed based on the previous pixel ends, the row index of the previous pixel is marked as the ending row index, and if the current pixel is not the last pixel, then continue traversing the next pixel in the current column; if the current pixel is the last pixel but not the last column, then continue traversing the next column; if the current pixel is the last pixel and not the last column, then the traversal ends. If the current pixel is not the first pixel and both the previous pixel and the previous pixel are unreachable locations 1, then if the current pixel is not the last pixel, then continue traversing the next pixel in the current column; if the current pixel is the last pixel but not the last column, then continue traversing the next column; if the current pixel is the last pixel and not the last column, then the traversal ends. After traversing each column, obtain the rectangular connected component to be updated for each column, which has a starting row index and an ending row index, and the starting row index and the ending row index are different. For the rectangular connected components to be updated, compare the rectangular connected components of the current column with the rectangular connected components of the previous column to determine whether the connected component update condition is met. If yes, update the rectangular connected component; otherwise, do not update it. Finally, obtain the set of rectangular connected components, with the current column being the second column to the last column in sequence.

6. The optimal full-coverage path planning method for corridors based on grid map connected domains as described in claim 1, characterized in that, In S3, the rule for partitioning the rectangular connected region is as follows: The compressed two-dimensional array map is traversed column by column, and the column index of each column is read in turn. In each column, each row of pixels is traversed from top to bottom. In the compressed two-dimensional array map, 0 represents empty land that can be reached, and 1 represents obstacles that are inaccessible. Analyze the current pixel in the current column: For the current column being the first column: if the current pixel is the first row pixel and is reachable (0), then a new rectangular connected component is constructed based on the current pixel, the current first row is marked as the starting row index of the newly constructed rectangular connected component, and the next pixel in the current column is traversed; if the current pixel is the first row pixel and is unreachable (1), then the first row cannot be used as the starting row index, and the next pixel in the current column is traversed directly. If the current pixel is not the first pixel in the row and both the previous pixel and the first pixel are reachable from the ground 0, then the range of the rectangular connected region built based on the previous pixel is extended to the current pixel. If the current pixel is not the last pixel in the row, then the next pixel in the current column is traversed. If the current pixel is the last pixel in the row, then the current rectangular connected region ends, the current last pixel is marked as the index of the endpoint row, and the next column is traversed. If the current pixel is not the first pixel in the row and is reachable from the ground 0, and the previous pixel is inaccessible from the ground 1, then a new rectangular connected region is built based on the current pixel, the current row is marked as the index of the starting row of the newly built rectangular connected region, if the current pixel is not the last pixel in the row, then the next pixel in the current column is traversed. If the current pixel is the last pixel in the row, the next column is traversed. If the current pixel is not the first pixel in the row and is an unreachable location 1, and the previous pixel is a reachable location 0, then the rectangular connected component constructed based on the previous pixel ends, and the row index of the previous pixel is marked as the ending row index. If the current pixel is not the last pixel in the row, then continue traversing the next pixel in the current column; if the current pixel is the last pixel in the row, then continue traversing the next column. If the current pixel is not the first pixel in the row and both the previous pixel and the previous pixel are unreachable locations 1, then continue traversing the next pixel in the current column if the current pixel is not the last pixel in the row; if the current pixel is the last pixel in the row, then continue traversing the next column. For cases where the current column is not the first column: if the current pixel is the first row pixel and is reachable (0), then a new rectangular connected component is constructed based on the current pixel, the current first row is marked as the starting row index of the newly constructed rectangular connected component, and the next pixel in the current column is traversed; if the current pixel is the first row pixel and is not reachable (1), then the first row cannot be used as the starting row index, and the next pixel in the current column is traversed directly. If the current pixel is not the first pixel in the row and both the previous pixel and the first pixel are reachable from the ground 0, then the range of the rectangular connected region built based on the previous pixel is extended to the current pixel. If the current pixel is not the last pixel in the row, then the next pixel in the current column is traversed. If the current pixel is the last pixel in the row, then the current rectangular connected region ends. The current last pixel is marked as the endpoint row index. The current column and the previous column are analyzed to see if they meet the connected region update condition. If they do, then update; otherwise, do not update and continue traversing the next column. If the current pixel is not the first pixel in the row and is reachable from the ground 0, and the previous pixel is inaccessible from the ground 1, then a new rectangular connected region is built based on the current pixel. The current row is marked as the starting row index of the newly built rectangular connected region. If the current pixel is not the last pixel in the row, then the next pixel in the current column is traversed. If the current pixel is the last pixel in the row, the current column and the previous column are analyzed to see if they meet the connected region update condition. If they do, then update; otherwise, do not update and continue traversing the next column. If the current pixel is not the first pixel and is an unreachable location 1, and the previous pixel is a reachable location 0, then the rectangular connected component constructed based on the previous pixel ends, and the row index of the previous pixel is marked as the ending row index. If the current pixel is not the last pixel, then continue traversing the next pixel in the current column. If the current pixel is the last pixel, then analyze whether the current column and the previous column satisfy the connected component update condition. If yes, update; otherwise, do not update, and continue traversing the next column. If the current pixel is not the first pixel and both the previous pixel and the previous pixel are unreachable locations 1, then continue traversing the next pixel in the current column. If the current pixel is the last pixel, then analyze whether the current column and the previous column satisfy the connected component update condition. If yes, update; otherwise, do not update, and continue traversing the next column. After traversing each column, we obtain a set of rectangular connected regions that have been partitioned and updated, with a starting row index and an ending row index that are different from each other.

7. The optimal full-coverage path planning method for connected domain corridors based on grid maps as described in claim 5 or 6, characterized in that, Connectivity update condition: If the starting row index and ending row index of the rectangular connected component in the current column are exactly the same as the starting row index and ending row index of the rectangular connected component in the previous column, then the rectangular connected component in the current column and the rectangular connected component in the previous column will be merged into one rectangular connected component.

8. The optimal full-coverage path planning method for corridors based on grid map connected domains as described in claim 1, characterized in that, S1 includes: S11. Perform grayscale processing on the two-dimensional raster map, including removing outliers and smoothing obstacle edges; S12. Expand the edges of obstacles in the processed map according to the set expansion radius; S13. Generate a fine-grained binary grayscale raster map.

9. A grid map-based optimal full-coverage path planning system for connected domain corridors, characterized in that, include: A map preprocessing module: acquires a two-dimensional raster map of the lawn environment, and performs image processing on the two-dimensional raster map to generate a fine-grained binary grayscale raster map. A map scaling adjustment module: Based on adaptive resolution adjustment rules, the coordinate system of the binary grayscale raster map is compressed and adjusted in two directions to generate a compressed two-dimensional array map; A rectangular connected component partitioning module: Based on the rectangular connected component partitioning rules, the compressed two-dimensional array map is analyzed by traversing column by column, and the set of rectangular connected components is dynamically generated and updated. The rule for partitioning the rectangular connected components is as follows: The compressed two-dimensional array map is traversed column by column, and the column index of each column is read in turn. In each column, each row of pixels is traversed from top to bottom, and rectangular connected components are divided based on the pixels. After the column-by-column traversal and division, the adjacent rectangular connected components are merged and updated to dynamically generate a set of rectangular connected components. Alternatively, during the column-by-column traversal and division process, the adjacent rectangular connected components are merged and updated to dynamically generate a set of rectangular connected components. An optimal full-coverage path planning module: For each rectangular connected region in the set of rectangular connected regions, the ox-plowing full-coverage path planning algorithm is used to generate the travel path in a single rectangular connected region. For the rectangular connected regions in the set of rectangular connected regions, the heuristic algorithm of A* is used to search for the path between the rectangular connected regions and generate the safe corridor between the rectangular connected regions, thereby constructing the optimal full-coverage planning path.

Images