Method and device for filtering sudden noisy points of automatic driving radar

Abstract

The invention discloses a sudden noise filtering method and device for an automatic driving radar, and the method comprises the steps: building a plurality of grids according to the current traveling speed of an automatic driving vehicle when the automatic driving radar carried by the automatic driving vehicle obtains the point cloud of a current frame; performing connected domain judgment on each grid by adopting the point cloud of the previous frame, and selecting a target grid according to a judgment result; according to the point cloud number of the current frame point cloud in each target grid, the burst noisy points are determined from the current frame point cloud and filtered, the filtering of the point cloud can be completed before obstacle classification through a noisy point screening mode of combining the grid with the connected domain, the noisy point filtering efficiency in a complex scene is effectively improved, and the automatic driving safety is guaranteed.

Description

Technical Field

[0001] This invention relates to the field of noise filtering technology, and in particular to a method and apparatus for filtering burst noise in autonomous driving radar. Background Technology

[0002] With the continuous development of autonomous driving technology, it places extremely high demands on the accuracy of environmental perception. LiDAR, because its ranging results directly reflect the distance in the real world and are not affected by scene factors, has become a core sensor to ensure driving safety, especially in complex road conditions where it plays a key role in safety redundancy.

[0003] However, in actual driving scenarios, LiDAR is easily affected by the diversity of the illumination surface. For example, scenarios such as colored road markings and reflections from standing water can easily produce multipath effects, leading to sudden noise in single or multiple frames of point cloud data. This type of noise is sudden and may persist, interfering with obstacle recognition, causing misjudgments by the autonomous driving system, and seriously threatening driving safety.

[0004] To address this, existing technologies filter noise by segmenting obstacles, pre-calculating the occupied grid based on tracking speed, and considering the temporal relationship between consecutive frames. While these methods can solve some sudden noise issues, they have significant limitations in practical applications. The tracking speed of obstacles needs to be calculated by fitting multi-frame point cloud data, resulting in a response lag in noise filtering. Furthermore, existing algorithms do not consider the continuous occurrence of noise and rely excessively on obstacle segmentation accuracy, which can easily lead to noise filtering failure in complex scenarios and reduce the safety of autonomous driving. Summary of the Invention

[0005] This invention provides a method and apparatus for filtering sudden noise in autonomous driving radar, which solves the technical problem that in the prior art, the tracking speed of obstacles needs to be calculated by fitting multi-frame point cloud data, and the noise filtering has a response lag. Furthermore, the existing algorithms do not consider the continuous occurrence of noise and rely too much on the accuracy of obstacle segmentation, which can easily lead to noise filtering failure in complex scenarios and reduce the safety of autonomous driving.

[0006] This invention provides a method for filtering burst noise in autonomous driving radar, comprising:

[0007] When the autonomous driving radar on the autonomous vehicle acquires the current frame point cloud, multiple grids are established according to the current travel speed of the autonomous vehicle.

[0008] The previous frame point cloud is used to determine the connectivity of each grid cell, and the target grid cell is selected according to the determination result.

[0009] Based on the number of point clouds in each target grid of the current frame point cloud, burst noise is determined from the current frame point cloud and filtered.

[0010] Optionally, when the autonomous driving radar on the autonomous vehicle acquires the current frame point cloud, establishing multiple grids according to the current travel speed of the autonomous vehicle includes:

[0011] When the autonomous driving radar on the autonomous vehicle acquires the point cloud of the current frame, the grid distance is obtained by multiplying the maximum speed of the autonomous vehicle with the frame rate interval of the autonomous driving radar.

[0012] The grid size is obtained by calculating the ratio between the current speed of the autonomous vehicle and the frame rate interval of the autonomous driving radar.

[0013] Using the location of the autonomous driving radar as the origin, multiple grids conforming to the grid size are established within the grid distance along the horizontal and vertical directions.

[0014] Optionally, the step of using the point cloud of the previous frame to determine the connected components of each of the grid cells, and selecting the target grid cell according to the determination result, includes:

[0015] Traverse the grid and determine whether any connected component of each grid is occupied by the point cloud of the previous frame;

[0016] If so, the grid is determined to be a non-target grid;

[0017] If not, then the grid is determined to be the target grid.

[0018] Optionally, determining and filtering burst noise from the current frame point cloud based on the number of point clouds in each of the target grids includes:

[0019] Count the number of point clouds in each of the target grids in the current frame;

[0020] If the number of point clouds is not zero, and the number of point clouds above the preset height threshold does not exceed the first obstacle threshold, and the number of point clouds does not exceed the second obstacle threshold, then the current frame point cloud within the target grid is determined as a burst noise.

[0021] Filter out the sudden noise.

[0022] Optionally, the method further includes:

[0023] If the number of point clouds exceeding a preset height threshold exceeds a first obstacle threshold, or if the number of point clouds exceeds a second obstacle threshold, then the current frame point cloud within the target grid is identified as a dynamic obstacle.

[0024] The present invention also provides a burst noise filtering device for autonomous driving radar, comprising:

[0025] A grid construction module is used to build multiple grids according to the current travel speed of the autonomous vehicle when the autonomous vehicle's radar acquires the current frame point cloud.

[0026] The target raster filtering module is used to determine the connectivity of each raster using the point cloud of the previous frame, and select the target raster according to the determination result.

[0027] The noise determination and filtering module is used to determine and filter burst noise from the current frame point cloud based on the number of point clouds in each of the target grids.

[0028] Optionally, the grid construction module is specifically used for:

[0029] When the autonomous driving radar on the autonomous vehicle acquires the point cloud of the current frame, the grid distance is obtained by multiplying the maximum speed of the autonomous vehicle with the frame rate interval of the autonomous driving radar.

[0030] The grid size is obtained by calculating the ratio between the current speed of the autonomous vehicle and the frame rate interval of the autonomous driving radar.

[0031] Using the location of the autonomous driving radar as the origin, multiple grids conforming to the grid size are established within the grid distance along the horizontal and vertical directions.

[0032] Optionally, the target grid filtering module is specifically used for:

[0033] Traverse the grid and determine whether any connected component of each grid is occupied by the point cloud of the previous frame;

[0034] If so, the grid is determined to be a non-target grid;

[0035] If not, then the grid is determined to be the target grid.

[0036] Optionally, the noise determination and filtering module is specifically used for:

[0037] Count the number of point clouds in each of the target grids in the current frame;

[0038] If the number of point clouds is not zero, and the number of point clouds above the preset height threshold does not exceed the first obstacle threshold, and the number of point clouds does not exceed the second obstacle threshold, then the current frame point cloud within the target grid is determined as a burst noise.

[0039] Filter out the sudden noise.

[0040] Optionally, the noise determination and filtering module is specifically used for:

[0041] If the number of point clouds exceeding a preset height threshold exceeds a first obstacle threshold, or if the number of point clouds exceeds a second obstacle threshold, then the current frame point cloud within the target grid is identified as a dynamic obstacle.

[0042] As can be seen from the above technical solutions, the present invention has the following advantages:

[0043] When the autonomous driving radar on board the autonomous vehicle acquires the current frame point cloud, multiple grids are established according to the current travel speed of the autonomous vehicle. The previous frame point cloud is used to determine the connectivity of each grid, and the target grid is selected according to the determination result. Based on the number of points in each target grid of the current frame point cloud, burst noise is identified from the current frame point cloud and filtered. By combining grids with connectivity to filter noise, the point cloud can be filtered before obstacle classification, which effectively improves the noise filtering efficiency in complex scenes and ensures the safety of autonomous driving. Attached Figure Description

[0044] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0045] Figure 1 A flowchart illustrating the steps of a burst noise filtering method for an autonomous driving radar provided in an embodiment of the present invention;

[0046] Figure 2 This is a schematic diagram of a grid construction method according to an embodiment of the present invention;

[0047] Figure 3 This is a structural block diagram of a burst noise filtering device for an autonomous driving radar provided in an embodiment of the present invention. Detailed Implementation

[0048] In existing technologies, the common approach is to segment the point cloud into obstacles and then filter them, pre-calculate the occupied grid based on the track speed of the obstacles, and finally filter based on the temporal relationship between consecutive frames. This algorithm can solve a large part of the sudden noise. However, this method has its own problems. First, it takes time for the obstacle to give its speed, usually at least three frames. Second, it ignores the possibility that sudden noise may be persistent, meaning it may not be a single frame.

[0049] To address this, embodiments of the present invention provide a method and apparatus for filtering sudden noise in autonomous driving radar, which solves the technical problem that in the prior art, the tracking speed of obstacles needs to be calculated by fitting multi-frame point cloud data, noise filtering has a response lag, and existing algorithms do not consider the continuous occurrence of noise and rely too much on obstacle segmentation accuracy, which can easily lead to noise filtering failure in complex scenarios and reduce the safety of autonomous driving.

[0050] To make the objectives, features, and advantages of this invention more apparent and understandable, the technical solutions of the embodiments of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the embodiments described below are only some embodiments of this invention, and not all embodiments. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0051] Please see Figure 1 , Figure 1 A flowchart illustrating the steps of a burst noise filtering method for an autonomous driving radar provided in an embodiment of the present invention.

[0052] This invention provides a method for filtering burst noise in autonomous driving radar, comprising:

[0053] Step 101: When the autonomous driving radar on the autonomous vehicle acquires the current frame point cloud, multiple grids are established according to the current travel speed of the autonomous vehicle.

[0054] A point cloud refers to a collection of discrete points in three-dimensional space generated by sensors such as LiDAR emitting laser beams and measuring information such as laser round-trip time and reflection intensity. Each point contains attributes such as spatial coordinates (e.g., x, y, z) and reflection intensity, which can be used to reconstruct the outline of the surrounding environment and is core data for autonomous driving to perceive obstacles and identify roads. In this embodiment, the current frame point cloud refers to the point cloud data collected and output by the autonomous driving radar at the current moment, used to reflect the latest environmental state around the vehicle in real time, and serves as an immediate input for perception and decision-making.

[0055] Autonomous driving radar refers to the active sensors carried by autonomous vehicles. By emitting and receiving lasers, it acquires three-dimensional point cloud data of the surrounding environment, providing environmental perception information for the autonomous driving system. Examples include lidar and LiDAR.

[0056] In this embodiment of the invention, point cloud data is collected frame by frame by the autonomous driving radar on the autonomous vehicle. Before the current frame point cloud is acquired and no obstacle is formed, multiple grids are established with the autonomous driving radar as the coordinate origin and combined with the current speed of the autonomous vehicle, so as to provide a data basis for subsequent filtering of sudden noise in the current frame point cloud.

[0057] In one example of the present invention, step 101 may include the following sub-steps:

[0058] When the autonomous driving radar on the autonomous vehicle acquires the point cloud of the current frame, the grid distance is obtained by multiplying the maximum speed of the autonomous vehicle with the frame rate interval of the autonomous driving radar.

[0059] The grid size is obtained by calculating the ratio between the current speed of the autonomous vehicle and the frame rate interval of the autonomous driving radar.

[0060] Using the location of the autonomous driving radar as the origin, multiple grids conforming to the grid size are established within the grid distance along the horizontal and vertical directions.

[0061] Existing technologies typically filter burst noise after obstacles form in the point cloud, which may lead to filtering delays. To improve filtering efficiency, when the current frame point cloud is acquired and no obstacles have formed in it, the product of the autonomous vehicle's maximum speed and the frame rate interval of the autonomous driving radar is calculated to obtain the grid distance. This ensures that the grid coverage can accommodate dynamic changes that may occur within the frame rate interval when the vehicle is traveling at high speed (such as obstacle movement or vehicle displacement), avoiding grid omissions due to insufficient speed estimation. Simultaneously, the ratio between the autonomous vehicle's current speed and the frame rate interval of the autonomous driving radar is calculated to obtain the required grid size.

[0062] After obtaining the grid size and grid distance, an origin grid is established with the location of the autonomous driving radar (grid 8) as the origin. Within this grid distance, multiple grids conforming to the grid size are established along both the horizontal and vertical directions. That is, grids are established with the radar origin, with a forward distance of A meters, totaling B grids, and a left and right distance of C meters, totaling D grids, for a total of B*D grids. Figure 2 As shown.

[0063] Step 102: Use the point cloud of the previous frame to determine the connectivity of each grid cell, and select the target grid cell according to the determination result;

[0064] The previous frame point cloud refers to the point cloud data output by the autonomous driving radar at the previous acquisition time. It records the environmental information at a slightly earlier time and can be temporally correlated with the current frame point cloud to help judge dynamic changes in the environment.

[0065] In this embodiment, after the grid construction is completed, the grid is used as the basic unit for connected component judgment. Connected component judgment is performed on each grid to determine whether the point cloud of the previous frame exists in any connected component. The target grid is selected from multiple grids according to the judgment result for subsequent judgment process.

[0066] In one example of the present invention, step 102 may include the following sub-steps:

[0067] Traverse the grid and determine whether any connected component of each grid is occupied by the point cloud of the previous frame.

[0068] If so, then the grid is determined to be a non-target grid;

[0069] If not, then the raster is determined to be the target raster.

[0070] In this embodiment of the invention, by traversing the grid, an eight-connected component judgment is performed on each grid to determine whether any connected component of each grid is occupied by the point cloud of the previous frame. If at least one connected component is occupied by the point cloud of the previous frame, since real obstacles usually persist in the time sequence frames and the point cloud distribution is continuous, it indicates that the grid may correspond to a real obstacle, and it is determined to be a non-target grid. If none of the connected components of the grid are occupied by the point cloud of the previous frame, it indicates that if it is occupied by the point cloud of the current frame, it may be a burst of noise, and it is determined to be a target grid.

[0071] Here, an eight-connected region refers to eight grid cells adjacent to a grid cell. In this embodiment, since these are all planar grid cells, then... Figure 2 As shown in grid 1, its eight connected regions are grids 2 to 9.

[0072] Step 103: Based on the number of point clouds in each target grid of the current frame point cloud, identify and filter burst noise from the current frame point cloud.

[0073] Point cloud count refers to the total number of point clouds within the target raster.

[0074] In this embodiment, after obtaining the target grid, the number of points in each target grid of the current frame point cloud is counted and compared with a preset obstacle threshold to determine whether the current frame point cloud is burst noise. If the current frame point cloud is determined to be burst noise, it is filtered out. The parts of the current frame point cloud that are not burst noise are left unprocessed.

[0075] In this embodiment, the autonomous driving radar continuously acquires the current frame point cloud and performs the above steps 101-103 for each frame point cloud to complete the real-time filtering of sudden noise. Since there is no need to judge the type of obstacle, it is more efficient than the prior art and can effectively ensure the smooth driving of autonomous vehicles.

[0076] In one example of the present invention, step 103 may include the following sub-steps:

[0077] Count the number of point clouds in each target grid of the current frame;

[0078] If the number of point clouds above the preset height threshold does not exceed the first obstacle threshold and the number of point clouds does not exceed the second obstacle threshold, then the current frame point cloud within the target grid is identified as burst noise.

[0079] Filter out sudden noise.

[0080] In this embodiment, the number of point clouds in each target grid of the current frame is counted. During driving, obstacles are usually higher than a preset height threshold. Static obstacles such as roadside vegetation are usually low in height and are often marked as open, so their point clouds are not counted in the above-mentioned point cloud count. Therefore, if the number of point clouds is not zero, and the number of point clouds higher than the preset height threshold does not exceed the first obstacle threshold and the number of point clouds does not exceed the second obstacle threshold, then the current frame point cloud in that target grid is identified as burst noise and this burst noise is removed.

[0081] In another example of the invention, the method further includes the following steps:

[0082] If the number of point clouds exceeding the preset height threshold exceeds the first obstacle threshold, or the number of point clouds exceeds the second obstacle threshold, then the current frame point cloud within the target grid will be identified as a dynamic obstacle.

[0083] In this embodiment of the invention, since dynamic obstacles can typically be of various types, such as pedestrians and vehicles, a first obstacle threshold is set to determine the number of point clouds exceeding a preset height threshold. This filters out static low-height point clouds such as the ground, curbs, and road markings, retaining "continuous objects with height," such as pedestrians and cars. Simultaneously, the number of point clouds is compared with a second obstacle threshold to determine if the number of point clouds within the target grid exceeds the second obstacle threshold. If so, it can be identified as a "large, low-lying object," such as a truck or construction vehicle, thereby reducing the risk of misjudgment.

[0084] The threshold for the second obstacle is greater than the threshold for the first obstacle.

[0085] Furthermore, for certain special scenarios, connected component determination can be performed using multiple frames. During this process, the first two and first three frames of point cloud data are used for connected component analysis. Only when a point cloud is consistently identified as a sudden noise point in multiple consecutive frames is it filtered out, preventing the accidental deletion of real dynamic objects, such as rapidly moving small obstacles. If the rate of change of the autonomous vehicle's speed exceeds a preset threshold, the grid size and grid distance can be recalculated to ensure that the grid coverage matches the vehicle's motion state, avoiding filtering errors caused by spatial partitioning failures.

[0086] In this embodiment of the invention, when the autonomous driving radar on the autonomous vehicle acquires the current frame point cloud, multiple grids are established according to the current travel speed of the autonomous vehicle; the previous frame point cloud is used to determine the connectivity of each grid, and a target grid is selected according to the determination result; based on the number of points in each target grid of the current frame point cloud, burst noise is determined from the current frame point cloud and filtered. By using a grid combined with connectivity noise filtering method, point cloud filtering can be completed before obstacle classification, effectively improving noise filtering efficiency in complex scenarios and ensuring the safety of autonomous driving.

[0087] Please see Figure 3 , Figure 3 A structural block diagram of a burst noise filtering device for an autonomous driving radar according to an embodiment of the present invention is shown.

[0088] This invention provides a burst noise filtering device for autonomous driving radar, comprising:

[0089] The grid construction module 301 is used to build multiple grids according to the current travel speed of the autonomous vehicle when the autonomous vehicle's autonomous driving radar acquires the current frame point cloud.

[0090] The target raster filtering module 302 is used to perform connectivity determination on each raster using the point cloud of the previous frame, and select the target raster according to the determination result;

[0091] The noise determination and filtering module 303 is used to determine and filter burst noise from the current frame point cloud based on the number of point clouds in each target grid of the current frame point cloud.

[0092] Optionally, the grid building module 301 is specifically used for:

[0093] When the autonomous driving radar on the autonomous vehicle acquires the point cloud of the current frame, the grid distance is obtained by multiplying the maximum speed of the autonomous vehicle with the frame rate interval of the autonomous driving radar.

[0094] The grid size is obtained by calculating the ratio between the current speed of the autonomous vehicle and the frame rate interval of the autonomous driving radar.

[0095] Using the location of the autonomous driving radar as the origin, multiple grids conforming to the grid size are established within the grid distance along the horizontal and vertical directions.

[0096] Optionally, the target grid filtering module 302 is specifically used for:

[0097] Traverse the grid and determine whether any connected component of each grid is occupied by the point cloud of the previous frame.

[0098] If so, then the grid is determined to be a non-target grid;

[0099] If not, then the raster is determined to be the target raster.

[0100] Optionally, the noise determination and filtering module 303 is specifically used for:

[0101] Count the number of point clouds in each target grid of the current frame;

[0102] If the number of point clouds is not zero, and the number of point clouds above the preset height threshold does not exceed the first obstacle threshold, and the number of point clouds does not exceed the second obstacle threshold, then the current frame point cloud within the target grid is identified as burst noise.

[0103] Filter out sudden noise.

[0104] Optionally, the noise determination and filtering module 303 is also specifically used for:

[0105] If the number of point clouds exceeding the preset height threshold exceeds the first obstacle threshold, or the number of point clouds exceeds the second obstacle threshold, then the current frame point cloud within the target grid will be identified as a dynamic obstacle.

[0106] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working process of the above-described device and module can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0107] In the several embodiments provided by this invention, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or modules may be electrical, mechanical, or other forms.

[0108] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical modules; that is, they may be located in one place or distributed across multiple network modules. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0109] Furthermore, the functional modules in the various embodiments of the present invention can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module. The integrated modules described above can be implemented in hardware or as software functional modules.

[0110] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for filtering burst noise in an autonomous driving radar, characterized in that, include: When the autonomous driving radar on the autonomous vehicle acquires the current frame point cloud, multiple grids are established according to the current travel speed of the autonomous vehicle. The previous frame point cloud is used to determine the connectivity of each grid cell, and the target grid cell is selected according to the determination result. Based on the number of point clouds in each target grid of the current frame point cloud, burst noise is determined from the current frame point cloud and filtered.

2. The method according to claim 1, characterized in that, When the autonomous driving radar on the autonomous vehicle acquires the current frame point cloud, multiple grids are established according to the current travel speed of the autonomous vehicle, including: When the autonomous driving radar on the autonomous vehicle acquires the point cloud of the current frame, the grid distance is obtained by multiplying the maximum speed of the autonomous vehicle with the frame rate interval of the autonomous driving radar. The grid size is obtained by calculating the ratio between the current speed of the autonomous vehicle and the frame rate interval of the autonomous driving radar. Using the location of the autonomous driving radar as the origin, multiple grids conforming to the grid size are established within the grid distance along the horizontal and vertical directions.

3. The method according to claim 1, characterized in that, The step of using the point cloud of the previous frame to determine the connectivity of each of the grid cells and selecting the target grid cell according to the determination result includes: Traverse the grid and determine whether any connected component of each grid is occupied by the point cloud of the previous frame; If so, the grid is determined to be a non-target grid; If not, then the grid is determined to be the target grid.

4. The method according to claim 1, characterized in that, The step of determining and filtering burst noise from the current frame point cloud based on the number of point clouds in each of the target grids includes: Count the number of point clouds in each of the target grids in the current frame; If the number of point clouds is not zero, and the number of point clouds above the preset height threshold does not exceed the first obstacle threshold, and the number of point clouds does not exceed the second obstacle threshold, then the current frame point cloud within the target grid is determined as a burst noise. Filter out the sudden noise.

5. The method according to any one of claims 1-4, characterized in that, The method further includes: If the number of point clouds exceeding a preset height threshold exceeds a first obstacle threshold, or if the number of point clouds exceeds a second obstacle threshold, then the current frame point cloud within the target grid is identified as a dynamic obstacle.

6. A burst noise filtering device for an autonomous driving radar, characterized in that, include: A grid construction module is used to build multiple grids according to the current travel speed of the autonomous vehicle when the autonomous vehicle's radar acquires the current frame point cloud. The target raster filtering module is used to determine the connectivity of each raster using the point cloud of the previous frame, and select the target raster according to the determination result. The noise determination and filtering module is used to determine and filter burst noise from the current frame point cloud based on the number of point clouds in each of the target grids.

7. The apparatus according to claim 6, characterized in that, The grid construction module is specifically used for: When the autonomous driving radar on the autonomous vehicle acquires the point cloud of the current frame, the grid distance is obtained by multiplying the maximum speed of the autonomous vehicle with the frame rate interval of the autonomous driving radar. The grid size is obtained by calculating the ratio between the current speed of the autonomous vehicle and the frame rate interval of the autonomous driving radar. Using the location of the autonomous driving radar as the origin, multiple grids conforming to the grid size are established within the grid distance along the horizontal and vertical directions.

8. The apparatus according to claim 6, characterized in that, The target grid filtering module is specifically used for: Traverse the grid and determine whether any connected component of each grid is occupied by the point cloud of the previous frame; If so, the grid is determined to be a non-target grid; If not, then the grid is determined to be the target grid.

9. The apparatus according to claim 6, characterized in that, The noise determination and filtering module is specifically used for: Count the number of point clouds in each of the target grids in the current frame; If the number of point clouds is not zero, and the number of point clouds above the preset height threshold does not exceed the first obstacle threshold, and the number of point clouds does not exceed the second obstacle threshold, then the current frame point cloud within the target grid is determined as a burst noise. Filter out the sudden noise.

10. The apparatus according to any one of claims 6-9, characterized in that, The noise determination and filtering module is specifically used for: If the number of point clouds exceeding a preset height threshold exceeds a first obstacle threshold, or if the number of point clouds exceeds a second obstacle threshold, then the current frame point cloud within the target grid is identified as a dynamic obstacle.

Images