Traversable area detection method and device, mobile tool, and storage medium

Abstract

The present invention discloses a method and device for detecting a traversable area. The method includes obtaining first parameter information and second parameter information, wherein the first parameter information includes point cloud data and the second parameter information includes positioning data of an intelligent mobile platform; determining the grid attributes of each grid in a grid map corresponding to a target area in the current frame based on the first parameter information and the second parameter information; determining the traversable area in the target area based on the grid attributes of each grid in the current frame, and outputting a target area map marked with the traversable area. The solution of the present invention uses a rule-based solution to detect traversable areas, while taking into account multiple frames of data, achieving higher coverage detection of traversable areas at a lower cost and avoiding missed detections; and can realize point cloud data processing and traversable area detection under a multi-radar combination, and has a very wide range of applications and is well suited for field environments such as unstructured roads.

Description

Technical Field

[0001] The present invention relates to the field of autonomous driving technology, and in particular to a traversable area detection method, a traversable area detection device, a mobile tool, and a storage medium. Background Art

[0002] Drivable area detection technology is a key component of environmental perception for autonomous vehicles, especially on semi-structured roads or unstructured roads in the wild. The presence of various amorphous obstacles such as potholes, ditches, puddles, cliffs, etc. can seriously affect the safety of autonomous vehicles. Therefore, detecting safe, drivable areas in the environment through drivable area detection technology is an important means to ensure vehicle operation safety.

[0003] Currently, there are two major categories of algorithms for detecting drivable areas in drivable area detection technology: the first category is detection solutions based on deep learning models; the second category is detection solutions based on rules. Among them, detection solutions based on deep learning models have the advantage of high accuracy, but they require a large amount of data to be labeled according to the operating scenario. At the same time, they have poor adaptability in unfamiliar environments and rely on parallel computing units in the computing platform (generally GPUs or other parallel processors), which requires high computing power. Rule-based detection solutions are based on the characteristics of each sensor, and manually design algorithms to extract the required features. The final calculation purpose is achieved through controllable algorithm design. Therefore, they have the advantages of low computational complexity and high real-time performance. They have received widespread attention in practical applications and have become an important research direction for drivable area detection.

[0004] Among rule-based detection solutions, those based on laser point clouds are more adaptable because they are unaffected by ambient lighting. The current mainstream solution based on laser point clouds is structured processing based on laser scan lines. This processing solution combines the relative relationships between points in adjacent laser scan line bundles, as well as the relationship between adjacent points on the same scan line, to determine longitudinal and transverse gradient changes. This solution has high sensitivity, but the disadvantage is that it requires the use of structured point clouds and is difficult to perform fusion calculations on multiple radar point clouds.

[0005] To overcome the various shortcomings of existing solutions, feature processing based on rasterization of laser point clouds is gradually entering the research field. This solution proposes a new solution approach: first rasterizing the laser point cloud and then extracting relevant features for traversability calculation. This approach can avoid the shortcomings and limitations of structured processing solutions based on laser scan lines. Currently, a proposed implementation of feature processing based on rasterization of laser point clouds is described in the invention patent publication number CN108873013A, entitled "A Method for Acquiring Road Passable Areas Using Multi-Line LiDAR." This method rasterizes the point cloud and sorts the points in each grid by height. Clustering is performed by setting a threshold based on the height arrangement to obtain multiple levels of plane blocks. The relative height of the plane blocks is then used to determine whether the grid belongs to the ground grid, obstacle grid, or hanging object grid. However, this solution only considers single-frame data, which can easily miss the ground between the laser beams, making it difficult to ensure the accuracy and coverage of the detection results. Furthermore, it has specific requirements for the type and scanning mode of the LiDAR, namely, it must be a multi-line LiDAR, which limits its applicable scenarios and poor universality. Summary of the Invention

[0006] An embodiment of the present invention provides a traversable area detection solution to solve the problems of the existing rule-based traversable area detection method only considering single-frame data and having specific requirements for the lidar, which easily leads to missing the ground between the laser beams and limited applicable scenarios.

[0007] In a first aspect, an embodiment of the present invention provides a method for detecting a traversable area, the method comprising:

[0008] Acquire first parameter information and second parameter information, wherein the first parameter information includes point cloud data, and the second parameter information includes positioning data of the intelligent mobile platform;

[0009] Determine the grid attributes of each grid in the grid map corresponding to the target area in the current frame according to the first parameter information and the second parameter information;

[0010] A passable area in the target area is determined according to the grid attributes of each grid in the current frame, and a target area map marked with the passable area is output.

[0011] In a second aspect, an embodiment of the present invention provides a traversable area detection device, which includes

[0012] A parameter acquisition module, configured to acquire first parameter information and second parameter information, wherein the first parameter information includes point cloud data and the second parameter information includes positioning data of the intelligent mobile platform;

[0013] a grid processing module, configured to determine a grid attribute of each grid in a grid map corresponding to a target area in a current frame according to the first parameter information and the second parameter information; and

[0014] The detection result determination module is used to determine the passable area in the target area according to the grid attributes of each grid in the current frame, and output a target area map marked with the passable area.

[0015] In a third aspect, an embodiment of the present invention provides another passable area detection device, comprising:

[0016] a memory for storing executable instructions; and

[0017] The processor is configured to execute the executable instructions stored in the memory, wherein the executable instructions implement the method steps of any embodiment of the present invention when executed by the processor.

[0018] In a fourth aspect, an embodiment of the present invention provides a mobile tool, which includes the passable area detection device described in the third aspect of the present invention.

[0019] In a fifth aspect, the present invention provides a storage medium having a computer program stored thereon, which implements the steps of the above method when executed by a processor.

[0020] The beneficial effects of the embodiments of the present invention are as follows: the method provided by the embodiments of the present invention adopts a rule-based scheme to detect passable areas, uses two input quantities, namely first parameter information and second parameter information, to perform point cloud rasterization and determine passable areas, and can directly output the determined passable areas. Compared with the scheme that can only determine the grid attributes based on a single frame of data, the detection results of the present invention have higher utilization value and more intuitive results; and the scheme of the embodiments of the present invention also uses the second parameter information including positioning data as the basis for determining the grid attributes, and takes into account multiple frames of data, thereby achieving higher coverage detection of passable areas at a lower cost, improving the accuracy of detection, and avoiding missed detection; in addition, the embodiments of the present invention also directly use point cloud data and positioning data as input quantities for detection, so there are no strict restrictions on the scanning mode and scanning radar of the laser radar, and can realize point cloud data processing and passable area detection under a multi-radar combination. It has a very wide range of applications and is well suited for field environments such as unstructured roads. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0022] Figure 1 Schematic diagram of a flow chart of a traversable area detection method according to an embodiment of the present invention;

[0023] Figure 2 Implementation of an embodiment of the present invention Figure 1 Flowchart of the method of step S11 in FIG.

[0024] Figure 3 Schematically shows the implementation of an embodiment of the present invention Figure 2 Flowchart of the method of step S21 in FIG.

[0025] Figure 4 Schematically illustrates the implementation of an embodiment of the present invention Figure 1 Flowchart of the method of step S12 in FIG.

[0026] Figure 5 Schematically illustrates the implementation of an embodiment of the present invention Figure 4 Flowchart of the method of step S40 in FIG.

[0027] Figure 6 A flowchart of a method for performing road attribute propagation calculation on a first seed grid according to an embodiment of the present invention is schematically shown;

[0028] Figure 7 Schematically shows a flow chart of a method for converting a corresponding second seed grid into a grid road according to an embodiment of the present invention;

[0029] Figure 8 Schematically illustrates another embodiment of the present invention. Figure 1 Flowchart of the method of step S12 in FIG.

[0030] Figure 9 Schematically shows the implementation of an embodiment of the present invention Figure 8 The method flow chart of step S42 in FIG.

[0031] Figure 10 The following schematically shows a principle block diagram of a passable area detection device according to an embodiment of the present invention;

[0032] Figure 11 This is a principle block diagram of a traversable area detection device according to another embodiment of the present invention;

[0033] Figure 12 This is a principle block diagram of a traversable area detection device according to another embodiment of the present invention;

[0034] Figure 13 This is a principle block diagram of a traversable area detection device according to another embodiment of the present invention;

[0035] Figure 14 This is a principle block diagram of a traversable area detection device according to another embodiment of the present invention;

[0036] Figure 15 This is a principle block diagram of a mobile tool according to one embodiment of the present invention;

[0037] Figure 16 Schematic diagram of the structure of an embodiment of the traversable area detection device of the present invention. DETAILED DESCRIPTION

[0038] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings 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 making creative efforts shall fall within the scope of protection of the present invention.

[0039] It should be noted that, unless there is any conflict, the embodiments and features in the embodiments of this application can be combined with each other.

[0040] The present invention may be described in the general context of computer-executable instructions, such as program modules, executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, and the like that perform specific tasks or implement specific abstract data types. The present invention may also be practiced in distributed computing environments where tasks are performed by remote processing devices connected via a communications network. In a distributed computing environment, program modules may be located in both local and remote computer storage media, including storage devices.

[0041] In the present invention, "module", "device", "system" and the like refer to related entities applied to a computer, such as hardware, a combination of hardware and software, software or software in execution, etc. Specifically, for example, an element can be, but is not limited to, a process running on a processor, a processor, an object, an executable element, an execution thread, a program and / or a computer. In addition, an application or script program running on a server, or a server can all be an element. One or more elements can be in an execution process and / or thread, and an element can be localized on a computer and / or distributed between two or more computers, and can be run by various computer-readable media. An element can also communicate through local and / or remote processes based on a signal having one or more data packets, for example, a signal from a data packet interacting with another element in a local system, a distributed system, and / or a signal from a network on the Internet that interacts with other systems via signals.

[0042] Finally, it should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "include" and "comprise" include not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article or device. In the absence of further limitations, the elements defined by the phrase "include..." do not exclude the presence of other identical elements in the process, method, article or device that includes the elements.

[0043] The passable area detection method in the embodiment of the present invention can be applied to a passable area detection device, so that a user or an intelligent mobile platform can use the passable area detection device to obtain a passable area, so as to control the intelligent mobile platform to perform corresponding driving actions according to the passable area. These passable area detection devices include, but are not limited to, detectors on autonomous driving vehicles, smart tablets, personal PCs, computers, cloud servers, etc. In particular, the passable area detection method in the embodiment of the present invention can also be applied to intelligent mobile platforms such as autonomous driving vehicles, unmanned cleaning vehicles, unmanned sweeping vehicles, sweeping robots, robots, semi-autonomous driving equipment, etc., and the present invention is not limited to this.

[0044] Figure 1 The diagram schematically shows a method for detecting a passable area according to an embodiment of the present invention. The execution subject of the method can be a detector or controller on an autonomous driving vehicle, or a processor of a passable area detection device such as a smart tablet, a personal PC, a computer, a cloud server, or a processor of an intelligent mobile platform such as an unmanned cleaning vehicle, an unmanned sweeping vehicle, a sweeping robot, an autonomous driving vehicle, a semi-autonomous driving device, etc. The embodiment of the present invention is not limited to this. The embodiment of the present invention is preferably described in detail by taking the execution subject as a processor on an intelligent mobile platform as an example. Figure 1 As shown, the method of the embodiment of the present invention includes:

[0045] Step S10: Acquire first parameter information and second parameter information, wherein the first parameter information includes point cloud data, and the second parameter information includes positioning data of the intelligent mobile platform;

[0046] Step S11: determining the grid attributes of each grid in the grid map corresponding to the target area in the current frame according to the first parameter information and the second parameter information;

[0047] Step S12: determining the passable area in the target area according to the grid attributes of each grid in the current frame, and outputting a target area map marked with the passable area.

[0048] In step S10, the acquired point cloud data may be point cloud data from a lidar sensor, or point cloud data from any sensor capable of generating similar 3D data, such as 3D point cloud data provided by an RGBD camera. Preferably, the point cloud data acquired in step S10 in this embodiment of the present invention does not need to be structured and is comprised of point cloud data from a lidar sensor, a binocular vision sensor such as an RGBD camera, or other sensor combinations capable of providing 3D point cloud data. For example, it may be comprised of point cloud data from one or more lidar sensors, or 3D point cloud data from different sensors. As a preferred implementation example, point cloud data refers to information data about the surrounding environment obtained by a binocular vision sensor such as a lidar sensor, an RGBD camera, or other sensor combinations capable of providing 3D point cloud data. It includes at least a series of three-dimensional coordinates (referred to as point cloud coordinates in this embodiment of the present invention) and may further include other information, such as auxiliary information such as reflectivity or color. It should also be noted that the point cloud data acquired in step S10 is a calibrated lidar point cloud or sensor 3D point cloud, meaning that the included point cloud coordinates are relative to the vehicle's origin. In this example, the self-driving vehicle is the intelligent mobile platform. The self-driving vehicle is the location of the sensors that acquire point cloud data. In other application scenarios, the self-driving vehicle can also refer to other intelligent mobile platforms, such as semi-autonomous driving devices, robots, and sweepers, where sensors acquire point cloud data.

[0049] As a preferred embodiment, the positioning data of the intelligent mobile platform includes positioning results of the absolute position and heading of the vehicle, which may be from a positioning module or other front modules (such as a perception module) on the intelligent mobile platform.

[0050] Preferably, the grid properties defined for each grid in the embodiment of the present invention include grid coordinates and grid status. Exemplarily, the embodiment of the present invention defines three basic states for each grid: road grid, obstacle grid and unknown grid. In particular, when defining and judging the grid properties, either only single-frame data or multiple frames of data can be considered at the same time. In order to achieve full coverage detection of the passable area in the environment through multiple frames of data and improve the coverage and accuracy of the detection, the preferred embodiment of the present invention simultaneously considers multiple frames of data to define and judge the grid properties. It should be noted that in the embodiment of the present invention, the concept of each frame is consistent with the unit and method of the perception module to perceive the data, that is, the perception module perceives one frame of perception data (such as image data) each time, and uses this as the basis for subsequent processing. Taking the perception data as the first parameter information and the second parameter information of the embodiment of the present invention as an example, the embodiment of the present invention obtains a frame of real-time point cloud data from the perception module and a frame of real-time positioning data from the positioning module as the input parameters of the current frame each time, and performs traversable area detection processing based on the input parameters of the current frame. The input parameters previously obtained will become the input parameters of the historical frame. Therefore, accordingly, the grid attributes of different frames can be determined based on the input parameters of different frames. The grid attribute record thus formed includes the grid attributes of each historical frame and the grid attributes of the current frame. The number of grid attributes of the historical frames included in the grid attribute record, that is, the number of historical frames retained, can be adjusted according to demand. For example, only the grid attributes of the next current frame before the current frame can be included as the grid attributes of the historical frame, or the grid attributes of the next current frame and the next next current frame before the current frame can be included as the grid attributes of the corresponding historical frame, or the grid attributes of all historical frames before the current frame can be included as the grid attributes of each historical frame. They are stored separately in the grid attribute record in units of frames or by frame ID. In other implementation examples, the data of the historical frames may also be accumulated, that is, the accumulated results of the historical frames are used as the grid attributes of the historical frames. This embodiment of the present invention is not limited to this.

[0051] Taking the example that the raster attribute record includes the raster attributes of each historical frame and the raster attributes of the current frame, that is, each frame corresponds to the raster attributes of the corresponding frame, the unknown raster in the embodiment of the present invention can be defined as a raster that has no sufficient number of point cloud inputs from the past to the present, that is, the number of point clouds corresponding to the raster in each historical frame and the current frame does not meet the first preset value and is determined to be an unknown raster; the road raster can be defined as a raster that has had a sufficient number of point clouds in the past or present and the point cloud distribution conforms to the road raster attribute, that is, in each historical frame and the current frame, as long as the number of point clouds corresponding to the raster in a certain frame meets the first preset value and the point cloud distribution meets the first preset condition, it is determined to be a road raster; the obstacle raster can be defined as a raster that has had point clouds exceeding the minimum threshold in the past or present and the point cloud distribution conforms to the obstacle raster attribute, that is, in each historical frame and the current frame, as long as the number of point clouds corresponding to the raster in a certain frame meets the first preset value and the point cloud distribution meets the second preset condition, it is determined to be an obstacle raster. Taking this as an example, Figure 2 The method process of implementing step S11 in one embodiment of the present invention is schematically shown. Figure 2 As shown, determining the grid attributes of each grid in the grid map corresponding to the target area in the current frame according to the first parameter information and the second parameter information can be specifically implemented by including the following steps:

[0052] Step S20: determining second parameter information corresponding to the first parameter information of the current frame according to the acquisition time of the first parameter information, and updating the grid coordinates of each grid in the current frame according to the determined second parameter information;

[0053] Step S21: determining the latest grid state corresponding to each grid after the grid coordinates are updated according to the first parameter information of the current frame;

[0054] Step S22: determining the grid state of each grid in the current frame according to the grid state of each grid before the grid coordinates are updated and the latest grid state.

[0055] In this embodiment of the present invention, step S20 is used to update the grid coordinates in the grid properties of the current frame, and steps S21 and S22 are used to update the grid state in the grid properties of the current frame. In other embodiments, after the grid properties of the current frame are updated, the grid properties of the historical frame are also saved to form a grid property record including the grid properties of the historical frame and the grid properties of the current frame. Of course, in other embodiments, the grid state and grid coordinates of the historical frame may also be saved when the grid state and grid coordinates of the current frame are updated, and the order of updating the grid state and grid coordinates may not be prioritized.

[0056] In a preferred embodiment of the present invention, the grid coordinate update in step S20 refers to updating the previous grid to the grid corresponding to the current vehicle position based on the vehicle positioning data in the second parameter information as the vehicle moves; and the grid state update in steps S21 and S22 refers to updating the grid state corresponding to the current frame based on the latest grid state determined by the historical frame and the current latest frame data. In actual applications, since the first parameter information comes from the perception data and the second parameter information comes from the positioning data, the perception and positioning modules can be driven and implemented by two completely different modules, and thus the acquisition processes of the first parameter information and the second parameter information can be relatively independent. Therefore, in order to ensure more efficient grid state updates, in a preferred embodiment of the present invention, when the positioning data and the perception data are acquired, before the grid state is updated based on the first parameter information, the grid position is first updated based on the positioning data corresponding to the time when the perception data was acquired, that is, the second parameter information corresponding to the time when the first parameter information was acquired. This ensures that the determination and update of the grid state are meaningful and efficient, and avoids invalid or erroneous grid state updates. Among them, the update of the grid positioning may include updating the grid coordinates of the current frame and associating and storing the grid coordinates of the historical frame to form grid coordinate records corresponding to each grid; in some embodiments, the update of the grid positioning may also refer only to the update of the grid coordinates of the current frame.

[0057] Specifically, as a preferred embodiment, in step S21, the correspondence between the point cloud and the grid can be determined, and the point cloud feature data of each grid can be generated according to the correspondence between the point cloud and the grid, and then the latest grid state of each grid can be determined according to the point cloud feature data of each grid. Figure 3 The method flow of this embodiment is schematically shown, as shown in FIG. Figure 3 As shown, it includes the following steps:

[0058] Step S30: determining the correspondence between the point cloud and the grid;

[0059] Step S31: generating point cloud feature data of each grid according to the correspondence between the point cloud and the grid;

[0060] Step S32: Determine the latest grid state of each grid based on the point cloud feature data of each grid.

[0061] In step S30, determining the correspondence between the point cloud and the grid can be achieved by traversing all point cloud data in the first parameter information. During the traversal process, the projection coordinates of each point cloud in the grid map corresponding to the target area are first determined based on the point cloud coordinates in each point cloud data and the preset grid parameters. The projection coordinates of each point cloud are then compared with the updated grid coordinates of each grid to match the grid corresponding to each point cloud. The grid parameters can be pre-set during the initialization phase. For example, they can include the grid size and total grid range set during the rasterization of the target area perceived by a lidar or sensor. Thus, the grid range to which the point cloud is projected can be determined based on the point cloud coordinates, the grid size, and the total grid range, thereby determining the projection coordinates of the point cloud in the grid map corresponding to the target area. The specific implementation process of projecting the point cloud onto the grid map corresponding to the target area and rasterizing the target area according to the grid parameters to form the grid map corresponding to the target area can be referred to in the prior art and will not be further described here.

[0062] In step S31, generating point cloud feature data for each grid based on the correspondence between the point cloud and the grid can be achieved by inputting the point cloud height information in the point cloud data into the grid corresponding to the point cloud, and after traversing all point clouds in the current frame, i.e., the first parameter information currently acquired, the point cloud height information input into each grid is statistically analyzed to obtain point cloud feature data for each grid. The specific content of the point cloud feature data can be set and statistically analyzed as needed. Exemplarily, the point cloud feature data can be implemented to include the number of point clouds corresponding to each grid and the statistical feature quantity corresponding to each grid. The statistical feature quantity mainly refers to statistical data that can be used to characterize the state of the grid. It can be obtained based on the point cloud height information input into each grid, or it can be obtained based on the number of point clouds corresponding to the grid and the point cloud height information input into the grid. Exemplarily, it can include but is not limited to grid height, grid minimum height, grid maximum height, grid variance, etc. Among them, the grid height refers to the average value of the point cloud height information input in each grid, that is, the accumulated point cloud height information input in each grid is divided by the number of point clouds input in the grid; the grid minimum height refers to the lowest point cloud height in the point cloud height information input in each grid, that is, the height of the point cloud with the lowest height in each grid is used as the lowest height of the grid; the grid maximum height refers to the maximum point cloud height in the point cloud height information input in each grid, that is, the height of the point cloud with the highest height in each grid is used as the maximum height of the grid; the grid variance refers to the variance of the point cloud distribution within each grid, which can be used to measure the discreteness of the point cloud corresponding to each grid within the grid.

[0063] In a preferred embodiment of the present invention, the point cloud feature data includes the number of point clouds corresponding to each grid in the current frame and statistical features corresponding to each grid. The statistical features include grid height, grid height difference, grid variance, and height difference from grids in the neighborhood that have been determined to be road grids. The number of point clouds corresponding to each grid in the current frame is obtained by counting the number of point clouds corresponding to each grid under the first parameter information of the current frame. The statistical features can be calculated based on the input point cloud height information for the current frame or by cumulatively counting the point cloud height information for all input historical frames and the current frame. For example, in step S32, determining the latest grid state of each grid based on the point cloud feature data of each grid can be implemented by traversing all grids in the target area and, during the traversal, determining the latest grid state of each grid based on the definition of the grid's basic state. For example, using the basic state definition described above as an example, during the traversal process, the number of point clouds in the current frame of the current grid can be determined for each grid. If the number of point clouds in the current frame does not reach a first preset value, the latest grid state of the corresponding grid is determined to be an unknown grid. If the number of point clouds in the current frame reaches the first preset value and the statistical feature quantity corresponding to the grid satisfies a first preset condition, the latest grid state of the corresponding grid is determined to be a road grid. If the number of point clouds in the current frame reaches the first preset value and the statistical feature quantity corresponding to the grid satisfies a second preset condition, the latest grid state of the corresponding grid is determined to be an obstacle grid. The first preset value, the first preset condition, and the second preset condition can be set and adjusted based on experience and needs. For example, the first preset condition can be set to: the grid height difference is less than a set threshold, the grid variance is less than a set threshold, and the height difference between the grid height and the height of a grid in the neighborhood that has been determined to be a road grid is less than a set threshold. The second preset condition can be set to: both the grid height difference and the grid variance exceed set thresholds.

[0064] In other embodiments, the latest grid state of each grid is determined based on the point cloud feature data of each grid, and the number of point clouds in the historical frame and the current frame and the statistical feature values ​​corresponding to each grid can also be simultaneously referenced. In this case, the point cloud feature data can also be implemented as including the number of point clouds in each historical frame and the current frame corresponding to each grid and the statistical feature values ​​corresponding to each grid. The number of point clouds in each historical frame and the current frame corresponding to each grid is obtained by counting the number of point clouds corresponding to each grid under the first parameter information of each frame, including the number of point clouds corresponding to each grid in the current frame and each previous historical frame. Taking the example of the point cloud feature data including the number of point clouds in each historical frame and the current frame corresponding to each grid and the statistical feature values ​​corresponding to each grid in the current frame, in this scenario, in step S32, determining the grid state of each grid in the current frame based on the point cloud feature data of each grid can be specifically implemented as follows: traversing all grids in the target area, and determining the grid state of each grid according to the definition of the basic grid state during the traversal process. Exemplarily, taking the basic state definition described above as an example, during the traversal process, the number of point clouds in each historical frame and the current frame of the current grid can be judged for each grid. When the number of point clouds in both the historical frame and the current frame does not reach a first preset value, the grid state of the corresponding grid is determined to be an unknown grid; when the number of point clouds in any frame among the number of point clouds in each historical frame and the current frame reaches a first preset value and the statistical feature quantity corresponding to the grid meets a first preset condition, the grid state of the corresponding grid is determined to be a road grid; when the number of point clouds in any frame among the number of point clouds in each historical frame and the current frame reaches a first preset value and the statistical feature quantity corresponding to the grid meets a second preset condition, the grid state of the corresponding grid is determined to be an obstacle grid.

[0065] As a preferred embodiment, since the first preset value in the embodiment of the present invention is a prerequisite for distinguishing unknown grids from other grid states, the grid size and point cloud density can be considered simultaneously when setting the first preset value to ensure high detection coverage. Preferably, the embodiment of the present invention sets the first preset value to be approximately 80% higher than the maximum number of noise points in the corresponding grid, that is, first preset value = maximum number of noise points in the grid + 80% * maximum number of noise points. For example, in the case of a grid size of 0.2x0.2 meters, the number of noise points generally does not exceed three. The maximum number of noise points for a grid of this size is 3. Accordingly, adding approximately 80% to the maximum number of noise points as the first preset value yields a first preset value of 5 for this grid size, which is more appropriate. This avoids setting the first preset value too small to be statistically significant, or too large to detect areas with sparse point clouds in a timely manner.

[0066] Since for an obstacle grid, a grid whose point cloud number reaches or exceeds the maximum number of noise points of the grid and whose point cloud distribution is not completely planar can be considered an obstacle grid, in other embodiments, the first preset value can also be set to be the same as the maximum number of noise points, that is, the first preset value is set to the maximum number of noise points of the grid as the minimum threshold standard for obstacle grid judgment, thereby more accurately detecting obstacle grids and avoiding missed detection of obstacle grids.

[0067] Accordingly, because the grid height difference, grid variance, and the height difference between the grid height and the grid height of nearby grids identified as road grids are highly correlated with the road grid attribute characteristics, road grids can be identified by judging these three values ​​when the grid point cloud is sufficiently large. Considering the correlation between these three values ​​and the road attribute grid, the larger the values, the less stringent the requirements for road roughness, and the easier it is to identify a road as a road grid. However, if these three values ​​are too large, low curbs or other small obstacles may be identified as roads, which could be dangerous. Therefore, in this embodiment of the present invention, a first preset condition is preferably set to judge these three values ​​to more accurately identify road grids. For example, the judgment thresholds for these three values ​​in the first preset condition can be set to be related to the grid size. Taking the grid size of 0.2x0.2 meters as an example, when paving the road surface, the grid height difference threshold in the first preset condition can be set to 0.2m, the grid variance threshold to 0.01m, and the height difference threshold between the grid height and the grid in the neighborhood that has been determined to be a road grid to be 0.1m, thereby improving the accuracy of road grid recognition. According to the grid states defined above, a grid's basic states are divided into three categories: unknown grid, road grid, and obstacle grid. Based on this, embodiments of the present invention preferably determine grids with insufficient point cloud counts as unknown grids. For grids with point cloud counts that meet a first preset value, obstacle grids and road grids are distinguished based on the first and second preset conditions. To ensure that a grid has only three states: unknown grid (insufficient number), road grid, and obstacle grid, the second preset condition can be set to the inverse of the first preset condition. That is, the second preset condition can be set to have the same threshold as the first preset condition, but with opposite judgment criteria. For example, the first preset condition is that each parameter is less than the threshold, while the second preset condition is that each parameter exceeds the threshold. This makes the obstacle grid and road grid states mutually exclusive. Consequently, the determined grid state can only be one of the three states: unknown grid (insufficient number), road grid, or obstacle grid, ensuring comprehensive coverage and accurate detection of different grid states.

[0068] After determining the latest grid state of each grid, as a preferred embodiment of the present invention, in step S22, the grid state of the current frame of each grid can be jointly determined based on the latest grid state of each grid and the grid state of the historical frame, and the grid state of the current frame of each grid can be updated. It should be noted that when the first parameter information and the second parameter information of the current frame are received, the grid properties before the current frame become the grid properties of the historical frame, that is, the grid state of the historical frame includes the grid state corresponding to the first parameter information and the second parameter information of at least one frame before the first parameter information and the second parameter information of the current frame.

[0069] In one embodiment of the present invention, the grid state of each grid in the current frame is determined based on the latest grid state of each grid and the grid state of the historical frame. Exemplarily, this can be achieved by comparing the latest grid state of each grid with the grid state before the grid coordinates of the grid are updated (i.e., before updating the grid state of the current frame, first compare the latest grid state obtained based on the parameter information of the current frame with the grid state corresponding to the parameter information of the frame before the current frame, and update the grid state of the current frame based on the comparison result), and select the optimal result as the grid state of the current frame. Exemplarily, taking the comparison of the latest grid state of each grid with the grid state of the grid before the grid coordinates are updated and selecting the optimal result as the grid state of the current frame as an example, the process of determining the grid state of the current frame will be described in detail below. Taking the current frame as the first parameter information and second parameter information of the nth frame, and the parameter information of the frame before the current frame as the first parameter information and second parameter information of the n-1th frame as an example, when the first parameter information and second parameter information of the current frame are received, the grid coordinates will be updated according to the first parameter information and second parameter information of the nth frame, and the latest grid state of each grid corresponding to the parameter information of the nth frame will be determined. When the first parameter information and second parameter information of the current frame are received, since each grid already has a grid coordinate and grid state determined based on the parameter information of the n-1th frame, the embodiment of the present invention preferably compares the latest grid state of each grid corresponding to the parameter information of the nth frame with the grid state of each grid corresponding to the parameter information of the n-1th frame, so as to select the optimal result as the grid state of the current frame, i.e., the nth frame, of the corresponding grid. When comparing the latest grid state of each grid corresponding to the parameter information of the nth frame with the grid state of each grid corresponding to the parameter information of the n-1th frame, the consistency of the comparison object will be guaranteed according to the update of the grid coordinates. For example, for grid x, its grid coordinates under the parameter information of the n-1th frame are (x n-1 ,y n-1 ), the grid coordinates under the parameter information of the nth frame are (x n ,y n), so when making comparisons, the grid coordinates under the parameter information of the n-1th frame are (x n-1 ,y n-1 The grid state of the grid (referred to as the grid state before the grid coordinates are updated in the embodiment of the present invention) is the same as the grid coordinates (x n ,y n ) is compared with the grid state of the grid in the current frame (referred to as the latest grid state after the grid coordinates are updated in the embodiment of the present invention) to ensure that the grid states of the same grid in different frames are compared to determine the grid state of the grid in the current frame. Among them, selecting the optimal result as the grid state of the current frame can be specifically implemented as follows: according to the comparison result, when the grid state before the grid coordinates are updated and the determined latest grid state are consistent, directly select the latest grid state as the grid state of the current frame, and update the grid state of the current frame of each grid to the determined latest grid state of the grid; and when the grid state before the grid coordinates are updated and the determined latest grid state are inconsistent, the grid state of the current frame is determined according to a preset strategy. Among them, the preset strategy can be set according to demand. Exemplarily, the preset strategy can be a strategy based on the lag comparison method, a strategy based on the Bayesian probability estimation method, a strategy based on the DS evidence theory method, etc., which is not limited by the embodiment of the present invention.

[0070] Taking the preset strategy of a hysteresis comparison method as an example, the hysteresis comparison method strategy can specifically determine the grid state of the current frame by selecting, from the latest grid state and the grid state before the grid coordinate update, the grid state that is more consistent with the development trend of the point cloud feature data of the corresponding grid as the grid state of the current frame. For example, taking the latest grid state as an obstacle grid and the grid state before the grid coordinate update as a road grid, based on the determination conditions for the obstacle grid and the road grid, if the point cloud feature data corresponding to the obstacle grid more closely meets the first preset value and the second preset condition, then the point cloud data of the grid is more consistent with the development trend of the obstacle grid state. In this case, the latest grid state, i.e., the obstacle grid, is selected as the grid state of the current frame. Among them, the grid state that is more consistent with the development trend of the point cloud feature data of the corresponding grid means that the point cloud feature data of the corresponding frame of the grid better meets the judgment condition of the corresponding grid state, that is, when the judgment condition of the grid state is greater than a certain parameter threshold, the point cloud feature data of the corresponding frame of the grid is greater than the threshold given in the judgment condition, and when the judgment condition of the grid state is less than a certain parameter threshold, the point cloud feature data of the corresponding frame of the grid is less than the threshold given in the judgment condition. For example, when the latest grid state is an obstacle grid and the grid state before the grid coordinate update is a road grid, if the difference between the point cloud feature data corresponding to the obstacle grid and the threshold parameter in the second preset condition is greater than the difference between the point cloud feature data of the road grid and the threshold parameter in the first preset condition, it means that the point cloud feature data of the obstacle grid is more consistent with the state trend of the grid. Therefore, the latest grid state whose point cloud data characteristics are more consistent with the grid state change trend will be selected as the grid state of the current frame. Taking the grid state defined above as an example, the threshold parameters in the first preset condition and the second preset condition can be a grid height difference threshold, a grid variance threshold, and a height difference threshold between the grid height and the grid in the neighborhood that has been determined to be a road grid. The threshold parameter of the first preset value can be the value of the first preset value.

[0071] Taking the preset strategy based on the Bayesian probability estimation method as an example, the strategy based on the Bayesian probability estimation method can specifically determine the grid state of the current frame. When determining the grid state of each grid in each frame, a passable attribute probability value is set for each grid based on the point cloud feature data of each grid in the corresponding frame. Therefore, the passable attribute probability value of each grid in each frame can be determined based on the first parameter information and the second parameter information obtained each time. After determining the latest grid state, the passable attribute probability value of each grid in the current frame is updated based on the Bayesian formula and the historical passable attribute probability values ​​of each grid in all frames up to the current frame. The grid state of the current frame is determined based on the updated passable attribute probability value. For example, if it is determined that the updated passable attribute probability value exceeds 50%, the grid state of the current frame of the grid is determined as a road grid. Otherwise, the grid state of the current frame of the grid is determined as an obstacle grid. In particular, for unknown grids, since there are not enough point clouds to make grid status judgments, the passable attribute probability value of the unknown grid can be assigned to 0 according to the number of point clouds. Therefore, when the passable attribute probability value of the grid is 0, the grid state of the current frame of the grid is determined to be an unknown grid.

[0072] Taking the preset strategy based on the DS evidence theory as an example, the strategy based on the DS evidence theory method to determine the grid state of the current frame can be specifically that when determining the grid state of each grid in each frame, different confidence levels are set for each grid according to the point cloud feature data of each grid in the corresponding frame. Thus, according to the first parameter information and the second parameter information obtained each time, the grid state of each grid in each frame and its corresponding confidence level can be determined. After determining the latest grid state, based on the latest grid state and its confidence level, as well as the grid state and its confidence level before the grid coordinates are updated, voting is performed through the set empirical formula to select the best result as the grid state of the current frame.

[0073] Therefore, the embodiment of the present invention can simultaneously use the point cloud data and grid status of the current frame and the historical frame to determine and update the grid status of the current frame, realize the use of historical information to supplement the grid between the point cloud bundles, and improve the coverage of the detection and judgment of the passable area.

[0074] In step S20, the new coordinates of the grid after the position change can be calculated based on the vehicle positioning data provided by the second parameter information, such as the absolute position and heading. The specific implementation method can be based on the projection relationship between the vehicle and the grid map and the selected coordinate system in the prior art, and will not be described in detail here. By calculating the new coordinates of each grid, the grid coordinates can be updated for each grid, so that the correspondence between the point cloud and the grid can be determined based on the new grid coordinates, the point cloud height information can be input to the grid, and the point cloud feature data of the grid can be statistically calculated.

[0075] In addition, after updating the grid state and grid coordinates of the grid (or before updating or at the same time), the embodiment of the present invention also saves the grid state of the historical frame (in some embodiments, it can also include the confidence level or passable attribute probability value corresponding to the grid state, etc.) and grid coordinates, so that the grid state of the current frame can be determined and updated in combination with the historical information. By referring to the historical information, the grids between the point cloud bundles can be supplemented, thereby improving the coverage and completeness of the passable area detection.

[0076] After the grid state and grid coordinates are updated, in step S12, embodiments of the present invention preferably determine the traversable area based on the updated grid state and grid coordinates. Specifically, the traversable area within the target area is preferably determined based on the grid attributes of each grid in the current frame. Specifically, this can be achieved by performing a road attribute propagation calculation in the grid map corresponding to the target area based on the grid state of each grid in the current frame.

[0077] Preferably, Figure 4 The specific processing flow of step S12 of a preferred embodiment of the present invention is schematically shown. Figure 4 As shown, step S12 can be implemented as including the following steps:

[0078] Step S40: performing road attribute propagation calculation in the grid map corresponding to the target area according to the grid status of each grid in the current frame, and marking the road grids in the target area;

[0079] Step S41: determining the passable area in the target area according to the marked road grid.

[0080] As a preferred embodiment, in step S40, the embodiment of the present invention performs road attribute propagation calculation by searching and traversing in the target area and confirming the grid status. Exemplarily, the search and traversal in the target area of ​​the embodiment of the present invention can be implemented as follows: setting the grid covered by the intelligent mobile platform (such as the self-driving vehicle in the case of an autonomous vehicle) as the initial grid, using the initial grid as a seed, traversing and searching other grids in the target area, judging the grid status in the current frame during the traversal search process, and searching and marking all road grids in the target area based on the judgment result. Among them, Figure 5 The specific implementation process of step S40 in one embodiment is schematically shown. Figure 5 As shown, step S40 can be specifically implemented as including the following steps:

[0081] Step S401: Initializing a road grid queue, wherein the initialized road grid queue includes at least one grid used as a propagation seed for road attribute propagation calculation;

[0082] Step S402: sequentially obtain a grid from the road grid queue as a first seed grid, perform road attribute propagation calculation on the currently obtained first seed grid according to the grid status of each grid in the current frame, and mark the road grid determined by the road attribute propagation calculation of each sequentially obtained first seed grid.

[0083] In step S401, the initialization of the road grid queue can be implemented, for example, by initializing the road grid queue based on the area covered by the intelligent mobile platform. For example, the grid covered by the intelligent mobile platform is set as the initial grid used as the propagation seed for the road attribute propagation calculation and placed in the road grid queue. In embodiments of the present invention, the road attribute propagation calculation preferably utilizes the grids in the formed road grid queue as seeds for the propagation of road grid attributes. Preferably, during initialization, setting the grid covered by the intelligent mobile platform as the initial grid used as the propagation seed for the road attribute propagation calculation specifically involves initializing the grid state of the grid covered by the intelligent mobile platform to road grids and setting the grid height of these grids to zero. Thus, after initialization, grids with road attributes are formed in the road grid queue. Since these grids are covered by the intelligent mobile platform, their road attributes are highly deterministic. Therefore, using these grids for the road attribute propagation calculation not only achieves comprehensive coverage of the target area but also makes the propagation of road attributes more reliable. It should be noted that the road grid queue in the embodiment of the present invention is the basis for performing road attribute propagation calculations. The grids stored therein are all grids that have road attributes. During the propagation calculation, road attributes can be propagated based on the road grids in the road grid queue. Therefore, in the specific road attribute propagation calculation process, the grids in the road grid queue can be used as propagation seeds.

[0084] In step S402, the road grid queue can be traversed to perform road attribute propagation calculations on each grid in the queue. Specifically, the following process can be implemented by looping until the grid road queue is empty: determining whether the road grid queue is empty. If the road grid queue is not empty, a grid is extracted as the first seed grid. The road attribute propagation calculation is then performed on the first seed grid in the grid state of the current frame, marking the road grid. If the road grid queue is determined to be empty, the traversal search process for the entire target area is terminated, thus completing the road attribute propagation calculation for the grid map corresponding to the target area.

[0085] The road attribute propagation calculation performed on the first seed grid can be implemented based on neighborhood search. Figure 6 The schematic diagram shows the specific implementation process of performing road attribute propagation calculation on the first seed grid and marking the road grid based on the neighborhood search method, refer to Figure 6 As shown, it can be implemented as including:

[0086] Step S601: determining a first group of propagation grids of a current first seed grid, and ending the road attribute propagation calculation for the current first seed grid when an obstacle grid exists in the first group of propagation grids according to the grid status of the first group of propagation grids;

[0087] Step S602: When there is no obstacle grid in the first set of propagation grids, the grid height of the unknown grid in the first set of propagation grids is marked as the height of the first seed grid, and road attribute propagation calculation is performed on each grid in the first set of propagation grids, and the road grid determined by the road attribute propagation calculation of each grid in the first set of propagation grids is marked.

[0088] In step S601, the first set of propagation grids determined for the current first seed grid is determined based on the principle of neighborhood search. For example, the first set of propagation grids can be determined as grids within a certain range around the first seed grid, such as grids within a radius R centered on the first seed grid. The surrounding certain range refers to the range of the neighborhood search for the first seed grid. The specific search range can be customized as needed and is not limited in this embodiment of the present invention. When determining whether an obstacle grid exists within the first set of propagation grids, the determination is made based on the grid status of each grid included in the first set of propagation grids in the current frame. In real-world scenarios, if an obstacle grid exists within a certain range of the neighborhood of a grid, it is difficult to form a traversable area. Therefore, when determining that an obstacle grid exists within a certain range around the first seed grid, the embodiment of the present invention stops propagating the road attributes of the current first seed grid, so that the ultimately formed traversable area is accurate and reliable.

[0089] More preferably, in practical applications, radars with smaller beams can only detect the properties of a very close area of ​​road, and the passable properties of farther areas cannot be determined due to blind spots between beams. Therefore, in order to achieve higher detection coverage and avoid missing detections in beam blind spots, if it is determined that no obstacle grid exists, when performing road propagation calculations, in addition to detecting and propagating the road grid, the embodiment of the present invention also propagates the ground height further through unknown grids with no obstacles nearby. Therefore, in step S602, when there is no obstacle grid in the first group of propagation grids, the grid height of the unknown grid in the first group of propagation grids is first marked as the height of the first seed grid, and then the road attribute propagation calculation is performed on all grids in the first group of propagation grids. Therefore, the embodiment of the present invention not only performs road propagation on road grids within the neighborhood of the first seed grid of road attributes, but also propagates ground height through unknown grids to propagate the grid's road attributes as far as possible, further expanding the overall detectable range. In this way, even radars with smaller beams can use road attribute propagation to detect and cover road attributes over a wider and more distant range. This reduces the requirements and reliance on the characteristics of the radar scanning beam, expands its applicability to a wider range of scenarios, and improves detection accuracy.

[0090] In one embodiment of the present invention, road attribute propagation calculations are performed on each grid in the first set of propagation grids. Marking the road grids determined by the road attribute propagation calculations for each grid in the first set of propagation grids can be specifically implemented by grid-roading each grid in the first set of propagation grids. This can specifically include: sequentially obtaining a grid from the first set of propagation grids as a second seed grid, determining a second set of propagation grids for each second seed grid, grid-roading the corresponding second seed grid based on the grid status of each grid in the second set of propagation grids for each second seed grid, and marking the second seed grid whose grid status is determined by grid-roading as a road grid. In a preferred embodiment, the second set of propagation grids determined for the second seed grids are the adjacent grids of each second seed grid. For example, in a rectangular grid, the adjacent grids of each grid are the four adjacent grids that share one edge with the grid. Grid-roading in this embodiment of the present invention refers to the process of confirming the grid status of the second seed grid as a road grid. Therefore, in the process of confirming the passable properties of the grid, the embodiment of the present invention not only considers the first parameter information and the second parameter information of the grid corresponding to each frame, but also considers the grid properties of the adjacent grids. Therefore, the determined grid state is more accurate, and the detection of the passable area can be achieved, and the detected passable area is more accurate and more reliable.

[0091] Figure 7The specific implementation process of performing grid roadization on the corresponding second seed grid to mark the second seed grid whose grid state is determined to be a road grid through grid roadization according to one embodiment of the present invention is schematically shown. Figure 7 As shown, it can be specifically implemented as including:

[0092] Step S701: Based on the grid status of each grid in the second set of propagation grids, determine whether there is a road grid in the second set of propagation grids. If so, proceed to step S702. Otherwise, that is, if there is no road grid in the second set of propagation grids, proceed to step S703.

[0093] Step S702: Confirming the road attributes of the corresponding second seed grid based on the road grid in the second set of propagation grids. When the corresponding second seed grid is determined to be a road grid, marking the grid corresponding to the second seed grid as a road grid in the grid map corresponding to the target area, and adding the second seed grid to the road grid queue.

[0094] Step S703: Determine whether there is an unknown grid marked with a grid height in the second set of propagation grids. If so, proceed to step S704. Otherwise, that is, if there is no unknown grid marked with a grid height, proceed to step S705.

[0095] Step S704: Confirming the road attributes of the corresponding second seed grid based on the unknown grid marked with grid height in the second set of propagation grids. If the corresponding second seed grid is determined to be a road grid, mark the grid corresponding to the second seed grid as a road grid in the grid map corresponding to the target area, and add the second seed grid to the road grid queue.

[0096] Step S705: Mark the current second subgrid as an obstacle grid.

[0097] In an embodiment of the present invention, if a road grid exists in the second set of propagation grids, it indicates that a road grid exists near the second seed grid. In this case, a more accurate nearby ground height is available for the current second seed grid. Therefore, the second seed grid can be grid-roadized based on the more accurate nearby ground height, thereby achieving the proximity propagation of the grid's road attributes. In this case, in step S702, when confirming the road attributes of the corresponding second seed grid based on the road grids in the second set of propagation grids, the road grids in the second set of propagation grids are used as the grids within the neighborhood of the second seed grid that have been determined to be road grids. Based on this, it is determined whether the second seed grid meets the first preset condition. If so, the second seed grid is considered to be a road grid, and the grid corresponding to the second seed grid is marked as a road grid in the grid map corresponding to the target area. The second seed grid is then placed in a road grid queue to be used as the first seed grid for road propagation calculation in subsequent searches. Specifically, when determining whether the second seed grid meets the first preset condition, the determination of whether the grid height difference is less than the set threshold and the grid variance is less than the set threshold is performed by comparing the grid height difference and the grid variance in the cloud feature data of the second seed grid with the set threshold. While the determination of whether the difference between the grid height and the grid height of a grid that has been determined to be a road grid in the neighborhood is less than the set threshold is performed by calculating the difference between the grid height of the second seed grid and the grid height of the road grid in the second group of propagation grids corresponding to the second seed grid, and comparing the difference with the set threshold.

[0098] If only the road attributes of the grid are propagated nearby, the propagation range and path may be easily interrupted due to scanning blind spots or insufficient point clouds. This means that long-distance propagation cannot be achieved. This can easily interrupt the search for the target area and the detection of the passable area, resulting in low coverage and lower detection efficiency and accuracy. Therefore, if it is determined that no road grid exists near the second seed grid, embodiments of the present invention preferably also implement jump propagation based on unknown grids, so that the road attributes of the grid can propagate further through the unknown grid, thereby achieving wider coverage and higher accuracy for the passable area detection method of embodiments of the present invention. Based on this, in step S703, if no road grid exists in the second set of propagated grids, embodiments of the present invention further determine whether there are unknown grids in the second set of propagated grids that are marked with grid heights. The marking of the heights of the unknown grids can be achieved through step S401 or step S602. That is, the unknown grids marked with grid heights have their road heights propagated through the road propagation calculation. Accordingly, in step S704, this embodiment of the present invention implements roadization of the second seed grid based on the road heights propagated from the unknown grid. Specifically, in step S704, the unknown grids in the second set of propagated grids, marked with grid heights, are used as grids within the neighborhood of the second seed grid that have been determined to be road grids. Based on these grid heights, a determination is made as to whether the second seed grid meets a first preset condition. If so, the second seed grid is considered a road grid, and the grid corresponding to the second seed grid is marked as a road grid in the grid map corresponding to the target area. The second seed grid is then placed in a road grid queue to be used as the first seed grid for road propagation calculations in subsequent searches. Specifically, when determining whether the second seed grid meets the first preset condition, the determination of whether the grid height difference is less than the set threshold and the grid variance is less than the set threshold is performed by comparing the grid height difference and the grid variance in the cloud feature data of the second seed grid with the set threshold. While the determination of whether the difference between the grid height and the grid height of the grid that has been determined to be a road grid in the neighborhood is less than the set threshold is performed by calculating the difference between the grid height of the second seed grid and the grid height of the unknown grid marked with the grid height in the second group of propagation grids corresponding to the determined second seed grid, and comparing the difference with the set threshold.

[0099] In step S705, if it is determined that there are no road grids near the second seed grid and no unknown grids marked with grid heights, then even if the second seed grid meets the first preset condition, it cannot be determined as a road grid and is directly determined as an obstacle grid. Thus, the method of the embodiment of the present invention uses the grids covered by the vehicle area as the starting point and performs road attribute propagation calculations in the target area based on the seed grids in the road grid queue and the grid states of each grid in the current frame to find the road grid. This propagation process always begins in the vehicle coverage area and includes both proximity propagation mediated by the road grid and jump propagation mediated by the unknown grid. As a result, the propagation distance is longer and the coverage area is wider. It can fully cover the blind spots between adjacent LiDAR beams, significantly improving the coverage rate of traversable area detection. In addition, when performing road attribute propagation calculations, the embodiments of the present invention propagate road attributes through the grids within the neighborhood of the seed grid or within a certain range around it. Therefore, there is also a proximity limit on the marking of road grids and unknown grids. The distance of road attribute propagation can be limited by setting the range of the first and second sets of propagation grids to achieve the goal of ensuring that the road attribute propagation calculation can cover adjacent lidar beams as the preferred target.

[0100] It should be noted that when the number of road grids or unknown grids marked with grid heights determined in steps S701 and S703 is more than one, the road attributes of the second seed grid can be determined in steps S702 and S704 based on the average grid height of the determined road grids or unknown grids marked with grid heights, or based on the grid height of any determined road grid or unknown grid marked with grid heights. The second seed grid can also be compared one by one with the grid heights of the determined road grids or unknown grids marked with grid heights, and the second seed grid can be determined as a road grid only when the grid height difference with each road grid or unknown grid marked with grid heights satisfies a first preset condition. This embodiment of the present invention is not limited in this regard. In addition, it can be understood that when the first preset condition is judged on the second seed grid, the second preset condition can also be judged by using the second seed grid's own grid height (when the second seed grid is a road grid) or the marked grid height (when the second seed grid is an unknown grid marked with a grid height).

[0101] After performing a traversal search and the aforementioned road attribute propagation calculation within the target area, embodiments of the present invention determine all road grids within the target area and mark all road grids in the grid map corresponding to the target area. In step S41, the traversable area within the target area is determined based on all marked road grids in the grid map corresponding to the target area. The determined traversable area is the minimum connected area defined by the road grids, i.e., the traversable area is the minimum connected area formed by connecting road grids. In a preferred embodiment, in step S41, the minimum connected area of ​​the road grid where the vehicle is located is preferably marked and output as the traversable area. Exemplarily, marking the traversable area can include drawing a boundary for the traversable area and marking the connected area enclosed by the boundary with information identifying the traversable road attributes, such as by color or text.

[0102] Figure 8 The specific processing flow of step S12 of another preferred embodiment of the present invention is schematically shown. Figure 8 As shown, step S12 can be implemented as Figure 4 The embodiment shown further includes the following steps:

[0103] Step S42: performing boundary correction on the determined passable area, and adjusting the passable area in the target area according to the boundary correction result.

[0104] In step S42, the embodiment of the present invention can filter the determined passable area boundary by combining preset initialization parameters, such as vehicle length, width, and turning radius, to delete small connected parts in the passable area and obtain a corrected passable area. For example, the vehicle width preset parameter set during initialization is used to filter the determined passable area boundary. Figure 9 The specific processing flow of step S42 of a preferred embodiment of the present invention is schematically shown as follows: Figure 9 As shown, step S42 can be implemented as including the following steps:

[0105] Step S421: generating a temporary binary raster map corresponding to the target area according to the determined passable area, wherein the temporary binary raster map is marked with the determined passable area and the area outside the determined passable area is marked as an impassable area;

[0106] Step S422: filtering the temporary binary grid image according to the pre-generated circular filter, and correcting the grid state of each grid in the temporary binary grid image;

[0107] Step S423: adjusting the determined passable area in the target area according to the corrected grid state in the temporary binary grid map.

[0108] The binary raster map in step S421 refers to a raster map containing only two characteristic values. In the embodiment of the present invention, the two characteristic values ​​are the passable attribute and the impassable attribute. Taking the example of generating a temporary binary raster map corresponding to the target area based on the passable area, the passable attribute is reflected by the grid state of the road grid, while the impassable attribute is reflected by the grid state outside the road grid. Therefore, the temporary binary raster map generated in step S421 refers to a temporary raster map corresponding to the target area, in which all grids within the passable area are set as road grids, and all grids outside the passable area are set as obstacle grids. In this way, the determined passable area is marked in the temporary binary raster map, and all areas outside the passable area are marked as impassable areas. It should be noted that the temporary binary raster map generated in the embodiment of the present invention is used for filtering to correct the boundaries of the passable area, and is different from the target area map that is finally output and marked with the passable area.

[0109] In step S422, the filter radius may be initialized based on a preset vehicle width to obtain a pre-generated circular filter. Preferably, the filter radius may be set to be equal to the preset vehicle width, and a circular filter with a radius R is generated based on the set filter radius R. For example, assuming that the preset vehicle width during the initialization process is one meter, the filter radius R is set to 1 meter, thereby forming a circular filter. The method of filtering the temporary binary raster image using the circular filter and correcting the passable area boundary in the binary raster image may preferably include: traversing all grids in the temporary binary raster image one by one using the circular filter, determining the grid status of all grids within the circular filter range during each traversal, and correcting the grid status of the grid corresponding to the current circular filter location based on the determination result. After the traversal is completed, the passable area boundary in the temporary binary raster image is re-determined based on the corrected grid status of all grids in the temporary binary raster image. For example, during each traversal, the position of the circular filter refers to the position of the grid that the circular filter is currently traversing. For example, when the circular filter is at position (x, y), it corresponds to the grid with the position (x, y) of the grid being traversed. All grids within the circular filter range refer to all grids within the coverage area of ​​the circular filter. According to the judgment result, the grid state of the grid corresponding to the current position of the circular filter can be corrected as follows: taking the coordinate position of the circular filter as (x, y) as an example, if the circular filter is at position (x, y), it is determined that all grids within the range of the circular filter are road grids, then the grid state of the grid at position (x, y) is corrected and marked as a road grid in the temporary binary grid map, otherwise it is corrected and marked as an obstacle grid. Thus, after the traversal is completed, a filtered temporary binary grid map can be generated, and the grid states of all grids are corrected in the filtered temporary binary grid map.

[0110] In step S423, the embodiment of the present invention re-determines all connected areas defined by the revised road grid corresponding to the target area based on the revised grid state in the temporary binary raster map. The connected area defined by the revised road grid where the vehicle is located is selected as the final revised traversable area. That is, the connected area defined by the road grid where the vehicle is located is selected as the determined traversable area. In step S12, the output target area map marked with traversable areas contains the revised traversable areas. In a preferred embodiment, the target area map output in step S12 is also a raster map corresponding to the target area, preferably a binary raster map marked with traversable and impassable areas. That is, in step S12, based on the grid state determined in step S423 or the aforementioned road attribute propagation calculation process, road grid areas in the binary raster map corresponding to the target area are marked as traversable areas, and other areas are marked as impassable areas. The binary raster map marked with traversable areas is then output. The output traversable area may be simply the connected domain formed by the road grid where the vehicle is located. In other embodiments, the generated and output target area map marked with traversable areas may also be a real-time map or a three-value grid map in which the locations corresponding to the road grid are marked as traversable areas. Of course, it is also possible to mark the traversable areas in the grid map formed by rasterizing the target area to output a grid map marked with traversable areas, etc. The present invention is not limited to this, as long as the traversable areas in the target area can be clearly marked and displayed. Therefore, the present invention can further filter the boundaries of the determined traversable areas, delete small connected parts in the traversable areas, and ensure the rationality and accuracy of the determined traversable areas.

[0111] As a preferred embodiment, in step S40, when performing road attribute propagation calculations on grids in the target area, a small portion of unknown grids is skipped. For example, the search range is dynamically adjusted based on changes in the searched grid status, skipping grid regions with unknown grid status to ensure that road grid attributes can be smoothly propagated to the target location. Specifically, when performing road attribute propagation calculations on other grids in the target area, such as when performing road attribute propagation calculations on the first seed grid, road propagation is performed by skipping unknown grids over short distances. More preferably, in an embodiment of the present invention, when there are no obstacle grids in the first set of propagation grids, before marking the grid heights of unknown grids in the first set of propagation grids as the heights of the first seed grid and performing road attribute propagation calculations on each grid in the first set of propagation grids, that is, before propagating the grid heights of the road grids to the unknown grids, the first set of propagation grids is re-determined by skipping unknown grids over a preset distance.

[0112] When acquiring point cloud data, depending on the installation location of different radar or sensor sensing modules and the point cloud data acquisition principle, the acquired point cloud data often contains blind spots. For example, consider a line-beam scanning LiDAR, a commonly used radar in the autonomous driving field. Its scanning principle is that each scanning line is typically a circular area, with different beams forming a certain angle between them. This means that the blind spot is located between the beam angles. Therefore, for this type of LiDAR, the fewer beams and the larger the beam angle, the larger the blind spot between the beams. Therefore, it is not difficult to understand that when performing road attribute propagation calculations within a blind spot, it is possible that the road attributes cannot be propagated to the next scan line due to the large blind spot and the presence of unknown cells within the blind spot. Therefore, as a preferred implementation, embodiments of the present invention allow for short-distance skipping of unknown cells during the neighborhood search during the road propagation calculation process. To ensure calculation accuracy, the range of unknown cells allowed to be skipped is preferably selected to be equivalent to the blind spot. For example, continuing with the lidar example, if a minimally configured 16-line radar is used as the sole sensing sensor, then in a typical installation, with only the road ahead, the first scan line in front of the vehicle is approximately eight meters from the front of the vehicle, and the second scan line is approximately ten meters from the front of the vehicle. Setting the grid size to be larger than one meter ensures that point clouds fall within each grid cell between eight and ten meters. However, a one-meter grid size is actually too large, as a relatively small vehicle is only 0.8 meters wide. A grid that is too large will prevent the vehicle from traversing narrow roads. Therefore, in actual applications, the grid size used may be as small as 0.2 meters. With such a small grid size, multiple grid cells between eight and ten meters will inevitably fail to receive any point clouds. Therefore, according to the aforementioned grid state definition conditions, after screening the first preset value, the grids at eight and ten meters will meet the condition of having a sufficient number of point clouds. However, the grids in the middle, that is, the grids in the blind spot range, will be determined as unknown grids because they do not meet the condition of having a sufficient number of point clouds. In this case, if road grid detection and propagation calculation are performed according to the road attribute propagation calculation method that strictly requires that the grids in the neighborhood must meet the road grid attributes, when performing a neighborhood search for the second seed grid to realize the grid roadization stage of the second seed grid, the search method for each grid in the first set of propagation grids requires that there is a road grid in the second set of propagation grids that is closely adjacent to the second seed grid and the height difference is less than the set threshold before road propagation can be performed. Therefore, when the grid at eight meters (the location of the first scan line) has been determined to be a road, it is no longer possible to cross the blind spot and extend the road attribute to ten meters.Therefore, as a preferred embodiment, in this scenario, it is possible to set the neighborhood search of the first seed grid, i.e., when determining the first set of propagation grids, based on the distance between two adjacent scan lines and the blind area covered by the distance, to allow the search range to be determined as skipping two meters of unknown grids and performing road height propagation (i.e., marking the grid height) on the unknown grids, so that the road attributes of the grid can be propagated from eight meters to the second scan line at ten meters. Specifically, the specific implementation method of skipping unknown grids over a short distance can be based on a preset skip distance. When it is determined that there are no obstacle grids in the first set of propagation grids, the grid height of the unknown grid in the first set of propagation grids is marked as the height of the first seed grid, and before the road attribute propagation calculation is performed on each grid in the first set of propagation grids, the unknown grids within a preset range of the preset distance from the first seed grid are determined as the first set of propagation grids to change the range of the neighborhood search. This can further prevent the inability to expand the road detection range due to the long-distance distribution of the laser beam, further improve the coverage of the passable area detection, and achieve a more complete passable area detection at a lower computational cost.

[0113] It should be noted that, in specific applications, the content of the above-mentioned statistical feature quantities can be flexibly adjusted according to needs, and the above-mentioned embodiments of the present invention are not regarded as limitations on their content. In specific implementations, they can be obtained by statistics on all grids or on columnar grids, as long as the statistical feature quantities and the number of point clouds can be used to accurately judge the grid status.

[0114] This invention uses a rule-based approach to rasterize point clouds and then propagates the grid's road attributes, starting with the vehicle itself. When updating grid states and propagating grid road attributes, it considers changes in the vehicle's positioning data and the grid state within the neighborhood. It also utilizes historical accumulated information based on changes in the vehicle's positioning data to supplement the grids between point cloud bundles, achieving a relatively complete detection of traversable areas at a low computational cost. Based on this approach, the invention imposes no strict restrictions on the scanning method of lidar (or sensor) or the combination of multiple lidars (or multiple sensors or a combination of lidars and sensors). It can be applied to a wide range of field environments, such as unstructured roads.

[0115] Figure 10 The present invention schematically shows a traversable area detection device according to an embodiment of the present invention, which can be applied to intelligent mobile platforms such as automatic / semi-automatic driving vehicles, unmanned sweepers, robots, etc. Figure 10 As shown, the device includes

[0116] A parameter acquisition module 50 is configured to acquire first parameter information and second parameter information, wherein the first parameter information includes point cloud data and the second parameter information includes positioning data of the intelligent mobile platform;

[0117] A grid processing module 51 is configured to determine a grid attribute of each grid in a grid map corresponding to a target area in a current frame according to the first parameter information and the second parameter information; and

[0118] The detection result determination module 52 is configured to determine the passable area in the target area according to the grid attributes of each grid in the current frame, and output a target area map marked with the passable area.

[0119] Figure 11 The following schematically shows a traversable area detection device according to a preferred embodiment of the present invention, as shown in FIG. Figure 11 As shown, in the preferred implementation example, the grid attributes are set to include grid coordinates and grid status, and the grid processing module 51 specifically includes:

[0120] A coordinate updating unit 510 is configured to determine second parameter information corresponding to the first parameter information of the current frame according to the acquisition time of the first parameter information, and update the grid coordinates of each grid in the current frame according to the determined second parameter information;

[0121] A state determining unit 511 is configured to determine the latest grid state corresponding to each grid after the grid coordinates are updated according to the first parameter information of the current frame; and

[0122] The state updating unit 512 is configured to determine the grid state of each grid in the current frame according to the grid state of each grid before the grid coordinates are updated and the latest grid state.

[0123] Figure 12 The following schematically shows a traversable area detection device according to a preferred embodiment of the present invention, as shown in FIG. Figure 12 As shown, the grid states include road grid, obstacle grid and unknown grid, and the detection result determination module 52 includes:

[0124] A road propagation unit 520 is configured to perform road attribute propagation calculation in the grid map corresponding to the target area according to the grid status of each grid in the current frame, and mark the road grids in the target area;

[0125] The area determination unit 521 is configured to determine a passable area in the target area according to the marked road grid.

[0126] Figure 13 The following schematically shows another preferred embodiment of a traversable area detection device of the present invention, as shown in FIG. Figure 13As shown, the detection result determination module 52 also includes:

[0127] The area correction unit 522 is configured to, after determining the passable area in the target area according to the marked road grid, further perform boundary correction on the determined passable area, and adjust the determined passable area in the target area according to the boundary correction result.

[0128] In a preferred embodiment, detection result determination module 52 specifically uses a road attribute propagation calculation method to search and traverse the grids within the target area, identifying road grids within the target area as grids with traversable attributes, and thereby determining the traversable area. The specific implementation and processing procedures of the various modules and units of the traversable area detection device in the above-mentioned embodiment of the present invention can be found in the description of the method section above, and will not be repeated here.

[0129] Figure 14 A traversable area detection device according to another embodiment of the present invention is schematically shown. As shown in the figure, the device includes:

[0130] Memory 60, for storing executable instructions; and

[0131] The processor 61 is configured to execute executable instructions stored in the memory, wherein the executable instructions, when executed by the processor, implement the steps of the passable area detection method described in any one of the aforementioned embodiments.

[0132] In specific practice, the above-mentioned traversable area detection device can be used, for example, on autonomous driving equipment or semi-autonomous driving equipment such as autonomous driving vehicles, unmanned cleaners, unmanned sweepers, robots, etc., to realize traversable area detection in the driving areas of these equipment, so as to improve the comfort of semi-autonomous driving or unmanned driving. Specifically, the above-mentioned traversable area detection device can be realized as a controller, planner, or detector on an intelligent mobile platform such as an autonomous driving vehicle.

[0133] Figure 15 A mobile tool according to an embodiment of the present invention is schematically shown. Figure 15 As shown, the mobile tool includes the passable area detection device 70 according to any of the above embodiments, so that the mobile tool can use the passable area detection device set thereon to detect the passable area, and then perform subsequent control such as direction, acceleration, throttle, brake, etc. based on the determined passable area.

[0134] Optionally, in actual applications, the mobile tool may also include a perception and recognition module and other planning and control modules, such as a path planning controller, a bottom-level controller, etc. The function of the passable area detection device 70 may also be implemented in the perception and recognition module or the planner, etc., and the embodiment of the present invention does not limit this.

[0135] The "mobile tool" referred to in the embodiment of the present invention can be a vehicle with L0-L5 autonomous driving technology levels established by the Society of Automotive Engineers International (SAE International) or the Chinese national standard "Automotive Driving Automation Classification".

[0136] For example, the mobile tool may be a vehicle device or a robotic device having the following various functions:

[0137] (1) Passenger-carrying function, such as family cars and buses;

[0138] (2) Cargo carrying function, such as ordinary trucks, box trucks, trailer trucks, closed trucks, tank trucks, flatbed trucks, container trucks, dump trucks, special structure trucks, etc.;

[0139] (3) Tool functions, such as logistics delivery vehicles, automated guided vehicles (AGVs), patrol cars, cranes, hoists, excavators, bulldozers, forklifts, rollers, loaders, off-road engineering vehicles, armored engineering vehicles, sewage treatment vehicles, sanitation vehicles, vacuum trucks, floor scrubbers, sprinkler trucks, sweeping robots, food delivery robots, shopping guide robots, lawn mowers, golf carts, etc.;

[0140] (4) Entertainment functions, such as entertainment vehicles, amusement park self-driving devices, balance vehicles, etc.;

[0141] (5) Special rescue functions, such as fire trucks, ambulances, power repair trucks, engineering rescue trucks, etc.

[0142] In some embodiments, an embodiment of the present invention provides a non-volatile computer-readable storage medium, which stores one or more programs including execution instructions, and the execution instructions can be read and executed by an electronic device (including but not limited to a computer, a server, or a network device, etc.) to execute the passable area detection method of any of the above embodiments of the present invention.

[0143] In some embodiments, an embodiment of the present invention further provides a computer program product, which includes a computer program stored on a non-volatile computer-readable storage medium, and the computer program includes program instructions. When the program instructions are executed by a computer, the computer executes the passable area detection method of any one of the above embodiments.

[0144] In some embodiments, an embodiment of the present invention further provides an electronic device comprising: at least one processor, and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the passable area detection method of any of the above embodiments.

[0145] In some embodiments, an embodiment of the present invention further provides a storage medium having a computer program stored thereon, which, when executed by a processor, implements the passable area detection method of any of the above embodiments.

[0146] Figure 16 FIG. 1 is a schematic diagram of the hardware structure of a passable area detection device provided by another embodiment of the present invention. The passable area detection device can be implemented with the structure shown in the figure, as shown in FIG. Figure 16 As shown, the passable area detection device includes:

[0147] One or more processors 610 and memory 620, Figure 16 A processor 610 is taken as an example.

[0148] The passable area detection device may further include: an input device 630 and an output device 640 .

[0149] The processor 610, the memory 620, the input device 630 and the output device 640 may be connected via a bus or other means. Figure 16 The bus connection is taken as an example.

[0150] Memory 620, as a non-volatile computer-readable storage medium, can be used to store non-volatile software programs, non-volatile computer-executable programs, and modules, such as the program instructions / modules corresponding to the traversable area detection method in the embodiments of the present invention. Processor 610 executes the non-volatile software programs, instructions, and modules stored in memory 620 to execute various server functional applications and data processing, thereby implementing the traversable area detection method in the above-described method embodiment.

[0151] The memory 620 may include a program storage area and a data storage area, wherein the program storage area may store an operating system and application programs required for at least one function; the data storage area may store data created based on the use of the passable area detection method, etc. In addition, the memory 620 may include a high-speed random access memory and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other non-volatile solid-state storage device. In some embodiments, the memory 620 may optionally include a memory remotely located relative to the processor 610, and these remote memories may be connected to the electronic device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, the Internet of Vehicles, an enterprise intranet, a local area network, a mobile communication network, and combinations thereof.

[0152] The input device 630 may receive input digital or character information and generate signals related to user settings and function control of the image processing device. The output device 640 may include a display device such as a display screen.

[0153] The one or more modules are stored in the memory 620 and, when executed by the one or more processors 610 , perform the passable area detection method in any of the above method embodiments.

[0154] The above-mentioned product can execute the method provided in the embodiment of this application, and has the functional modules and beneficial effects corresponding to the execution method. For technical details not fully described in this embodiment, please refer to the method provided in the embodiment of this application.

[0155] The electronic devices of the embodiments of the present application exist in various forms, including but not limited to:

[0156] (1) Mobile communication devices: These devices are characterized by their mobile communication capabilities and are primarily designed to provide voice and data communications. These terminals include smartphones (e.g., iPhones), multimedia phones, feature phones, and low-end phones.

[0157] (2) Ultra-mobile personal computer devices: These devices fall under the category of personal computers, have computing and processing capabilities, and generally also have mobile Internet access. These terminals include PDAs, MIDs, and UMPCs, such as the iPad.

[0158] (3) Portable entertainment devices: These devices can display and play multimedia content. These devices include audio and video players (such as iPods), handheld game consoles, e-books, smart toys, and portable car navigation devices.

[0159] (4) Server: A device that provides computing services. The server consists of a processor, hard disk, memory, system bus, etc. The server is similar to a general computer architecture, but because it needs to provide highly reliable services, it has higher requirements in terms of processing power, stability, reliability, security, scalability, and manageability.

[0160] (5) Other electronic devices with data interaction functions.

[0161] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of the modules may be selected based on actual needs to achieve the objectives of this embodiment.

[0162] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a general hardware platform, or of course, by hardware. Based on this understanding, the above technical solution, in essence, or the part that contributes to the relevant technology, can be embodied in the form of a software product. The computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a magnetic disk, an optical disk, etc., and includes a number of instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or certain parts of the embodiment.

[0163] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A traversable area detection method, characterized in that: The method comprises: Determine a passable area in the target area according to grid attributes of each grid in the current frame, and output a target area map marked with the passable area, wherein the grid attributes include grid states, and the grid states include road grids, obstacle grids, and unknown grids; The determining of the passable area in the target area according to the grid attributes of each grid in the current frame includes: Initializing and forming a road grid queue, wherein the initialized road grid queue includes at least one grid used as a propagation seed for road attribute propagation calculation; Sequentially obtain a grid from the road grid queue as a first seed grid, perform road attribute propagation calculation on the currently obtained first seed grid according to the grid status of each grid in the current frame, and mark the road grid determined by the road attribute propagation calculation of each sequentially obtained first seed grid; Determine the traversable area in the target area based on the marked road grid; The step of performing a road attribute propagation calculation on the currently acquired first seed grid according to the grid state of each grid in the current frame, and marking a road grid determined by the road attribute propagation calculation of each first seed grid acquired in sequence, includes: Determine a first group of propagation grids for the current first seed grid, and end the road attribute propagation calculation for the current first seed grid when an obstacle grid exists in the first group of propagation grids according to the grid status of the first group of propagation grids; When there are no obstacle grids in the first set of propagation grids, the grid height of the unknown grid in the first set of propagation grids is marked as the height of the first seed grid, and road attribute propagation calculation is performed on each grid in the first set of propagation grids, and the road grids determined by the road attribute propagation calculation for each grid in the first set of propagation grids are marked.

2. The method according to claim 1, characterized in that The grid properties of each grid in the current frame are determined in the following way: Acquire first parameter information and second parameter information, wherein the first parameter information includes point cloud data, and the second parameter information includes positioning data of the intelligent mobile platform; The grid attributes of each grid in the grid map corresponding to the target area in the current frame are determined according to the first parameter information and the second parameter information.

3. The method according to claim 2, characterized in that The grid attributes further include grid coordinates, and determining the grid attributes of each grid in the target area in the current frame according to the first parameter information and the second parameter information includes: Determining second parameter information corresponding to the first parameter information of the current frame according to the acquisition time of the first parameter information, and updating the grid coordinates of each grid in the current frame according to the determined second parameter information; Determine the latest grid state corresponding to each grid after the grid coordinates are updated according to the first parameter information of the current frame; The grid state of each grid in the current frame is determined according to the grid state of each grid before the grid coordinates are updated and the latest grid state.

4. The method according to claim 3, characterized in that Determining the latest grid state corresponding to each grid after the grid coordinates are updated according to the first parameter information of the current frame includes: Determining a correspondence between the point cloud data in the first parameter information of the current frame and each grid in the target area based on the first parameter information of the current frame and the updated grid coordinates of each grid, and generating point cloud feature data of each grid based on the correspondence between the point cloud data and each grid; The latest grid state of each grid is determined based on the point cloud feature data of each grid.

5. The method according to claim 1, wherein The method further comprises: performing road attribute propagation calculation on each grid in the first group of propagation grids, marking the road grids determined by the road attribute propagation calculation on each grid in the first group of propagation grids; A grid is sequentially obtained from the first group of propagation grids as a second seed grid, a second group of propagation grids is determined for each second seed grid, and the corresponding second seed grid is grid-roadized according to the grid state of each grid in the second group of propagation grids of each second seed grid. The second seed grid whose grid state is determined to be a road grid through grid roadization is marked.

6. The method according to claim 5, characterized in that The step of performing grid roadization on the corresponding second seed grid according to the grid status of each grid in the second group of propagation grids of each second seed grid, and marking the second seed grid whose grid status is determined to be a road grid through the grid roadization, includes: Determine whether there is a road grid in the second group of propagation grids based on the grid status of each grid in the second group of propagation grids; if so, confirm the road attribute of the corresponding second seed grid based on the road grid in the second group of propagation grids; if the corresponding second seed grid is determined to be a road grid, mark the grid corresponding to the second seed grid as a road grid in the grid map corresponding to the target area, and add the second seed grid to the road grid queue; If there is no road grid in the second set of propagation grids, determine whether there is an unknown grid marked with grid height in the second set of propagation grids. If so, confirm the road attribute of the corresponding second seed grid based on the unknown grid marked with grid height in the second set of propagation grids. When it is determined that the corresponding second seed grid is a road grid, mark the grid corresponding to the second seed grid as a road grid in the grid map corresponding to the target area, and add the second seed grid to the road grid queue.

7. The method according to claim 6, characterized in that When there is no obstacle grid in the first set of propagation grids, before marking the grid height of the unknown grid in the first set of propagation grids as the height of the first seed grid and performing road attribute propagation calculation on each grid in the first set of propagation grids, the method further includes: The first set of propagation grids is re-determined by skipping unknown grids of a preset distance.

8. The method according to claim 1, characterized in that The determining of the passable area in the target area according to the grid attributes of each grid in the current frame further includes: After determining the passable area in the target area according to the marked road grid, the boundary of the determined passable area is corrected, and the passable area in the target area is adjusted according to the boundary correction result.

9. The method according to claim 8, characterized in that Make boundary corrections to the identified passable areas, including: Generate a temporary binary raster map corresponding to the target area according to the determined passable area, wherein the temporary binary raster map is marked with the determined passable area and the area outside the determined passable area is marked as an impassable area; Filtering the temporary binary raster image according to a pre-generated circular filter, and correcting the grid state of each grid in the temporary binary raster image; The passable area in the target area is determined to be adjusted according to the modified grid state in the temporary binary grid map.

10. The passable area detection device is characterized in that: The device comprises: a detection result determination module, configured to determine a passable area in a target area based on grid attributes of each grid in a current frame, and output a target area map marked with the passable area, wherein the grid attributes include grid states, and the grid states include road grids, obstacle grids, and unknown grids; The determining of the passable area in the target area according to the grid attributes of each grid in the current frame includes: Initializing and forming a road grid queue, wherein the initialized road grid queue includes at least one grid used as a propagation seed for road attribute propagation calculation; Sequentially obtain a grid from the road grid queue as a first seed grid, perform road attribute propagation calculation on the currently obtained first seed grid according to the grid status of each grid in the current frame, and mark the road grid determined by the road attribute propagation calculation of each sequentially obtained first seed grid; Determine the traversable area in the target area based on the marked road grid; The step of performing a road attribute propagation calculation on the currently acquired first seed grid according to the grid state of each grid in the current frame, and marking a road grid determined by the road attribute propagation calculation of each first seed grid acquired in sequence, includes: Determine a first group of propagation grids for the current first seed grid, and end the road attribute propagation calculation for the current first seed grid when an obstacle grid exists in the first group of propagation grids according to the grid status of the first group of propagation grids; When there are no obstacle grids in the first set of propagation grids, the grid height of the unknown grid in the first set of propagation grids is marked as the height of the first seed grid, and road attribute propagation calculation is performed on each grid in the first set of propagation grids, and the road grids determined by the road attribute propagation calculation for each grid in the first set of propagation grids are marked.

11. The traversable area detection device according to claim 10, characterized in that: The device further comprises: A parameter acquisition module, configured to acquire first parameter information and second parameter information, wherein the first parameter information includes point cloud data and the second parameter information includes positioning data of the intelligent mobile platform; The grid processing module is used to determine the grid properties of each grid in the grid map corresponding to the target area in the current frame according to the first parameter information and the second parameter information.

12. The passable area detection device is characterized in that: include: a memory for storing executable instructions; as well as A processor, configured to execute executable instructions stored in a memory, wherein the executable instructions implement the steps of the method according to any one of claims 1 to 9 when executed by the processor.

13. A mobile tool, characterized in that The mobile tool includes: the passable area detection device as claimed in claim 12.

14. A storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of the method according to any one of claims 1 to 9 are implemented.

Images