Methods, devices, electronic equipment, and storage media for determining drivable areas

Abstract

This application discloses a method, apparatus, electronic device, and storage medium for determining a drivable area. The method includes: acquiring target point cloud data of a target area; wherein the target point cloud data is obtained by a detection device using target detection rays emitted by the device to detect the target area; mapping the target point cloud data onto each target detection ray of the detection device, and determining the distribution of data points on each target detection ray based on the data points mapped on each target detection ray; supplementing data points on each target detection ray using a data point supplementation strategy corresponding to the data point distribution; and determining the drivable area within the target area based on the data points on each supplemented target detection ray. Through the above method, this application can improve the accuracy of the determined drivable area.

Description

Technical Field

[0001] This application relates to the field of data processing technology, and in particular to a method, apparatus, electronic device, and storage medium for determining a drivable area. Background Technology

[0002] Currently, the detection of drivable areas for vehicles and robots can provide information on whether a certain area is drivable or not, thus providing input for the driving path of vehicles and robots. Among these, the accurate detection of drivable areas is an important research topic in the industry. Summary of the Invention

[0003] The main technical problem addressed by this application is to provide a method, apparatus, electronic device, and storage medium for determining a drivable area, which can improve the accuracy of the determined drivable area.

[0004] To address the aforementioned technical problems, this application provides a method for determining a drivable area. This method includes: acquiring target point cloud data of a target area; wherein the target point cloud data is obtained by a detection device using its emitted target detection rays to detect the target area; mapping the target point cloud data onto each target detection ray of the detection device, and determining the distribution of data points on each target detection ray based on the data points mapped onto each target detection ray; supplementing data points on each target detection ray using a data point supplementation strategy corresponding to the data point distribution; and determining the drivable area within the target area based on the data points on each supplemented target detection ray.

[0005] The data points include obstacle points and drivable area points; the distribution of data points includes at least one of the following: the first positional difference between two adjacent drivable area points on the target detection ray, the second positional difference between an obstacle point and an adjacent drivable area point on the target detection ray, the third positional difference between the nearest data point on the target detection ray and the target object, and the fourth positional difference between the first drivable area point on the target detection ray and the corresponding boundary point; wherein the detection device is installed on the target object, the nearest data point is the data point closest to the target object, the first drivable area point is the drivable area point farthest from the target object, and the corresponding boundary point is the boundary point of the target area corresponding to the target detection ray.

[0006] Specifically, based on the data points mapped on each target detection ray, the distribution of data points on each target detection ray is determined, including: for each target detection ray, obtaining the first positional distance between each second drivable area point on the target detection ray and its adjacent drivable area point to obtain the first positional difference; wherein, the second drivable area point is located before the nearest obstacle point on the target detection ray, and the corresponding adjacent drivable area point is located after the second drivable area point; and / or, for each target detection ray, obtaining the first positional distance between the nearest obstacle point on the target detection ray and its adjacent drivable area point. Two positional distances are used to obtain a second positional difference situation; wherein, the nearest obstacle point is the obstacle point closest to the target object, and the adjacent drivable area point corresponding to the nearest obstacle point is located in front of the nearest obstacle point; and / or, for each target detection ray, a third positional distance between the nearest data point on the target detection ray and the target object is obtained to obtain a third positional difference situation; and / or, for each target detection ray, a fourth positional distance between the first drivable area point on the target detection ray and the corresponding boundary point is obtained to obtain a fourth positional difference situation; wherein, there is no obstacle point between the first drivable area point and the corresponding boundary point.

[0007] The data point supplementation strategy, which utilizes the corresponding data point distribution, supplements data points on each target detection ray. This includes: for each target detection ray, supplementing data points in at least one interval region using different or the same data point supplementation strategy. The at least one interval region includes at least one of the following: a first interval region between two adjacent drivable area points with a difference greater than a first difference threshold; a second interval region between an obstacle point with a difference greater than a second difference threshold and an adjacent drivable area point; a third interval region between the nearest data point with a difference greater than a third difference threshold and the target object; and a fourth interval region between a first drivable area point with a difference greater than a fourth difference threshold and the corresponding boundary point.

[0008] The data point supplementation strategies corresponding to the distribution of each data point are the same; in at least one interval region, data points are supplemented using different or the same data point supplementation strategies, including: adding several supplementary drivable area points in at least one interval region; wherein the several supplementary drivable area points are set at equal intervals.

[0009] The data point distribution includes the second position difference situation; after supplementing the data points on each target detection ray using the corresponding data point distribution strategy, the method for determining the drivable area also includes: removing data points on the target detection ray that are located after the nearest obstacle point.

[0010] The data point distribution includes the fourth position difference situation; using the data point supplementation strategy corresponding to the data point distribution situation, data points are supplemented on each target detection ray, including: obtaining the angle between each target detection ray and the normal of the detection device, and using the detection rays with the angle within the range of the angle as reference detection rays; for each reference detection ray, data points are supplemented on each reference detection ray using the data point supplementation strategy corresponding to the fourth position difference situation.

[0011] The process of mapping target point cloud data onto the target detection rays of the detection device includes: mapping the target point cloud data onto a preset plane area to obtain several data points on the preset plane area; wherein the orthographic projection of the preset plane area on the ground is less than or equal to the orthographic projection of the target area on the ground, and the orthographic projection of the detection device on the preset plane area is within the preset plane area; for each data point, based on the mapping angle of the data point, mapping the data point onto the target detection ray that matches the mapping angle of the data point.

[0012] The target point cloud data includes non-ground point cloud data. Before mapping each data point to a target detection ray that matches the mapping angle of the data point, the method for determining the drivable area also includes: removing data points with heights greater than a height threshold from the obstacle point cloud data.

[0013] The process of determining drivable areas within the target region based on data points on the target detection ray after data supplementation includes: mapping the supplemented data points to a preset planar region; determining contour points of several drivable areas from the data points on the preset planar region; obtaining several connected regions based on the contour points of several drivable areas; wherein each connected region includes at least one contour point; and removing connected regions whose number of contour points does not meet the quantity requirement, and using the remaining connected regions as drivable areas within the target region.

[0014] The target point cloud data includes several three-dimensional points and the category of each three-dimensional point; acquiring the target point cloud data of the target area includes: acquiring several three-dimensional points collected by the detection device; performing category detection on each three-dimensional point to obtain the category of each three-dimensional point; wherein the category of the three-dimensional point includes at least the ground category and the obstacle category.

[0015] To address the aforementioned technical problems, another technical solution adopted in this application is: providing a device for determining a drivable area, the device comprising an acquisition module, a determination module, a supplementation module, and a generation module; the acquisition module is used to acquire target point cloud data of a target area; wherein, the target point cloud data is obtained by a detection device using its emitted target detection rays to detect the target area; the determination module is used to map the target point cloud data onto each target detection ray of the detection device, and determine the distribution of data points on each target detection ray based on the data points mapped on each target detection ray; the supplementation module is used to supplement data points on each target detection ray using a data point supplementation strategy corresponding to the data point distribution; the generation module is used to determine the drivable area in the target area based on the data points on each supplemented target detection ray.

[0016] To solve the above-mentioned technical problems, another technical solution adopted in this application is to provide an electronic device, which includes a processor and a memory, wherein the memory stores program instructions, and the processor executes the program instructions to implement the above-mentioned method for determining the drivable area.

[0017] To solve the above-mentioned technical problems, another technical solution adopted in this application is to provide a computer-readable storage medium for storing program instructions, which can be executed to implement the above-mentioned method for determining the drivable area.

[0018] The above technical solution compensates for data points based on the distribution of data points on each target detection ray, making the point cloud data more regular and complete, and solving the problems of data spikes and irregularities caused by the detection defects of the detection device. Therefore, the drivable area determined based on the data points on the target detection ray after data supplementation is more accurate and more consistent with reality. Attached Figure Description

[0019] Figure 1 This is a flowchart illustrating an embodiment of the method for determining a drivable area provided in this application;

[0020] Figure 2 This is a schematic diagram of an embodiment of the preset planar region provided in this application;

[0021] Figure 3 This is a schematic diagram of an embodiment of the first positional difference situation provided in this application;

[0022] Figure 4 This is a schematic diagram of an embodiment of the second positional difference situation provided in this application;

[0023] Figure 5 This is a schematic diagram of an embodiment of the third positional difference situation provided in this application;

[0024] Figure 6 This is a schematic diagram of an embodiment of the fourth positional difference situation provided in this application;

[0025] Figure 7 yes Figure 1 The flowchart of step S12 shown is a schematic diagram of one embodiment;

[0026] Figure 8 yes Figure 1 The flowchart of step S14 shown is a schematic diagram of one embodiment.

[0027] Figure 9 This is a schematic diagram of an embodiment of the device for determining the drivable area provided in this application;

[0028] Figure 10 This is a schematic diagram of the structure of an embodiment of the electronic device provided in this application;

[0029] Figure 11 This is a schematic diagram of an embodiment of the computer-readable storage medium provided in this application. Detailed Implementation

[0030] The embodiments of this application will now be described in detail with reference to the accompanying drawings.

[0031] In the following description, specific details such as particular system architectures, interfaces, and technologies are presented for illustrative purposes rather than for limiting purposes, in order to provide a thorough understanding of this application.

[0032] In this document, the term "and / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Additionally, the character " / " generally indicates that the preceding and following related objects have an "or" relationship. Furthermore, "many" in this document means two or more. Moreover, the term "at least one" in this document means any combination of at least two of any one or more of a plurality of objects. For example, including at least one of A, B, and C can mean including any one or more elements selected from the set consisting of A, B, and C.

[0033] Please see Figure 1 , Figure 1 This is a flowchart illustrating an embodiment of the method for determining a drivable area provided in this application. It should be noted that if substantially the same result is obtained, the embodiments of this application are not necessarily identical. Figure 1 The illustrated process sequence is limited. For example... Figure 1 As shown, this embodiment includes:

[0034] Step S11: Obtain target point cloud data for the target area.

[0035] For example, after converting a 3D non-ground point cloud into a mesh map, the nearest point at each angle is obtained using the bundle method, which is the boundary point of the drivable area. The resulting point set is very "rough," requiring noise filtering of the boundary points, followed by piecewise fitting and fusion. Finally, the polynomial of the continuous segments is resampled to form a uniform envelope of points, which represents the drivable area. However, because the boundary points are sometimes too "irregular," simply filtering noise without supplementation will result in significant errors during subsequent fitting, leading to a mismatch between the obtained area envelope and the actual area. Therefore, this embodiment supplements the acquired point cloud data to make the drivable area determined based on the supplemented point cloud data more accurate, thus improving the detection effect of the drivable area.

[0036] In this embodiment, target point cloud data of the target area is acquired; wherein, the target point cloud data is obtained by the detection device using the target detection rays emitted by it to detect the target area.

[0037] In one embodiment, the target point cloud data includes a number of three-dimensional points and the category of each three-dimensional point. Specifically, acquiring the target point cloud data of the target area involves: acquiring a number of three-dimensional points collected by the detection device; performing category detection on each three-dimensional point to obtain its category; wherein the category of the three-dimensional point includes at least drivable area and obstacle categories. That is, after acquiring the number of three-dimensional points detected by the detection device (including the three-dimensional coordinates x, y, z with the detection device as the origin and the reflection intensity, where the positive x-axis direction is the forward direction, the positive y-axis direction is the left side of the target object, and the positive z-axis direction is the vertical upward direction of the target object), the number of three-dimensional points are input into the detection network for category detection to determine the category of each three-dimensional point, i.e., determining whether each three-dimensional point is a drivable area point or an obstacle point.

[0038] In one specific implementation, drivable area points include ground points, sidewalk points, and lane line points, while obstacle points include fence points, vegetation points, traffic sign points, and vehicle (e.g., trucks, buses, cars, non-motorized vehicles, etc.) points.

[0039] In one specific embodiment, the detection device is a lidar. The lidar uses its emitted target detection rays to detect the target area and obtain target point cloud data of the target area. Specifically, the lidar is configured to detect a preset number of target detection rays in one circle (e.g., the preset number is 1080, and the resolution is...). A lidar system uses 1080 target detection rays to detect a target area and collect several three-dimensional points. The lidar obtains target point cloud data by detecting the target area using its emitted target detection rays, and stores the target point cloud data according to different angles, similar to forming target rays. Typically, it can be stored in the following format (the lidar's angular resolution is set to α): Angle 0, point cloud data queue: point 1, point 2, ..., point N1; Angle α, point cloud data queue: point 1, point 2, ..., point N2; Angle 2α, point cloud data queue: point 1, point 2, ..., point N3; ...; Angle π, point cloud data queue: point 1, point 2, ..., point Nm.

[0040] In one embodiment, the target point cloud data of the target area can be obtained from cloud storage or local storage. It should be noted that the target point cloud data is obtained by the detection device using its emitted target detection rays to detect the target area, and then stored in cloud storage or local storage. Of course, in other embodiments, the detection device can also obtain the target point cloud data of the target area in real time by using its emitted target detection rays.

[0041] The method for determining the drivable area provided in this application can be applied to the field of vehicles, such as the field of autonomous driving of vehicles, in which case the target area can be any area along the vehicle's route; of course, the method for determining the drivable area provided in this application can also be applied to the field of robots, such as the field of robotic vacuum cleaners, in which case the target area can be, but is not limited to, public places (such as parks, squares, offices, etc.) and residential areas (such as bedrooms, living rooms, kitchens, etc.).

[0042] Step S12: Map the target point cloud data onto each target detection ray of the detection device, and determine the distribution of data points on each target detection ray based on the data points mapped on each target detection ray.

[0043] In this embodiment, target point cloud data is mapped onto each target detection ray of the detection device. In one embodiment, such as Figure 2 As shown, Figure 2 This is a schematic diagram of an embodiment of the preset planar region provided in this application. A preset planar region can be set and a number of point rays can be laid on the preset planar region. Each point ray corresponds to a target detection ray with an angular resolution. Then, the data points mapped to the preset planar region are mapped to the corresponding point rays to map the target point cloud data to each target detection ray of the detection device.

[0044] Furthermore, in this embodiment, the distribution of data points on each target detection ray is determined based on the data points mapped onto each target detection ray. The distribution of data points on each target detection ray can characterize the detection process of the target ray within the target area; that is, it can characterize the missing detection information during the detection process. For example, due to limitations in the detection principle of the detection device itself, when detecting distant scenes, the collected three-dimensional points are missing or sparse. This is reflected on the target detection ray as missing data points at the end of the ray furthest from the target object, and sparse data points distributed at the end of the ray furthest from the target object.

[0045] In one embodiment, the data points include obstacle points and drivable area points; the distribution of data points includes at least one of the following: a first positional difference between two adjacent drivable area points on the target detection ray, a second positional difference between an obstacle point and an adjacent drivable area point on the target detection ray, a third positional difference between the nearest data point on the target detection ray and the target object, and a fourth positional difference between the first drivable area point on the target detection ray and its corresponding boundary point; wherein the detection device is disposed on the target object, the nearest data point is the data point closest to the target object, the first drivable area point is the drivable area point farthest from the target object, and the corresponding boundary point is the boundary point of the target area corresponding to the target detection ray. The type of target object is not limited; for example, the target object is a vehicle, a robot, etc.

[0046] In other embodiments, when the target point cloud data is mapped onto the target detection rays of the detection device through a preset planar area and several point rays laid on the preset planar area, such as Figure 2 As shown, the corresponding boundary points of each target detection ray are the intersection points of each target detection ray and the preset planar area.

[0047] Due to limitations in the detection principle of the detection device itself, there is significant information loss within the same angular range; that is, information detected by the target detection ray will be incomplete. The distribution of data points on the target detection ray includes the initial positional difference between two adjacent drivable area points on the target detection ray. By determining the positional difference between two adjacent drivable area points on the target detection ray, it is possible to determine whether there is information loss between two adjacent drivable area points, so that data points can be supplemented between adjacent drivable area points where information loss exists.

[0048] Due to limitations in the detection principle of the detection device itself, there is significant information loss within the same angular range; that is, information detected by the target detection ray will be incomplete. The distribution of data points on the target detection ray includes the second positional difference between obstacle points and adjacent drivable area points on the target detection ray. By determining the positional difference between obstacle points and their adjacent drivable area points on the target detection ray, it is possible to determine whether there is information loss between obstacle points and their adjacent drivable area points, so as to subsequently supplement data points between obstacle points and their adjacent drivable area points where information loss exists.

[0049] Within a certain distance of the target object on which the detection device is installed, the detection device cannot detect, forming a detection blind zone. The distribution of data points on the target detection ray includes the third positional difference between the nearest data point on the target detection ray and the target object. By determining the positional difference between the nearest data point on the target detection ray and the target object—that is, by determining the positional difference between the nearest data point that the detection device can detect and the target object—it is determined whether data points need to be added to the detection blind zone.

[0050] Due to reflection intensity issues, drivable area points far from the detection device are essentially undetectable, or rather, cannot be accurately detected. The data point distribution on the target detection ray includes the fourth positional difference between the first drivable area point and its corresponding boundary point on the target detection ray. By determining the positional difference between the drivable area point farthest from the target object and its corresponding boundary point on the target detection ray, it is possible to determine whether there is missing information between the drivable area point farthest from the target object and its corresponding boundary point. This allows for subsequent data supplementation between the drivable area points farthest from the target object and their corresponding boundary points where information is missing.

[0051] In one specific implementation, for each target detection ray, the first positional distance between each second drivable area point on the target detection ray and its adjacent drivable area point is obtained to obtain a first positional difference. The second drivable area point is located on the target detection ray before the obstacle point closest to the target object, and the corresponding adjacent drivable area point is located after the second drivable area point. Since some areas on the target detection ray after the obstacle point closest to the target object are necessarily not drivable areas of the target object, the first positional difference is obtained only by performing relevant calculations on the drivable area points before the obstacle point closest to the target object.

[0052] For example, such as Figure 3 As shown, Figure 3 This is a schematic diagram of an embodiment of the first positional difference situation provided in this application. Figure 3 Point A in the diagram represents a second drivable area point on the target detection ray. Figure 3 Point B is the adjacent drivable area point corresponding to point A. The first positional distance between points A and B on the target detection ray is obtained.

[0053] In one specific implementation, for each target detection ray, a second positional distance between the nearest obstacle point and its adjacent drivable area point on the target detection ray is obtained to determine a second positional difference. The nearest obstacle point is the obstacle closest to the target object, and the adjacent drivable area point corresponding to the nearest obstacle point is located before the nearest obstacle point. Since some areas after the nearest obstacle point on the target detection ray are necessarily not drivable areas of the target object, the second positional difference is obtained only by performing correlation calculations between the nearest obstacle point and its adjacent drivable area points.

[0054] For example, such as Figure 4 As shown, Figure 4 This is a schematic diagram of an embodiment of the second positional difference situation provided in this application. Figure 4 Point B in the diagram represents the obstacle point on the target detection ray that is closest to the target object. Figure 4 Point A is a drivable area point adjacent to point B. The second positional distance between points A and B on the target detection ray is obtained.

[0055] In one specific implementation, for each target detection ray, the third positional distance between the nearest data point on the target detection ray and the target object is obtained to determine the third positional difference. For example, such as... Figure 5 As shown, Figure 5 This is a schematic diagram of an embodiment of the third positional difference situation provided in this application. Figure 5 Point A in (a) is the closest drivable area point on the target detection ray to the target object. Figure 5 In (a), point O is the location of the target object. The third position distance between point A and point O on the target detection ray is obtained. Figure 5 Point A in (b) is the obstacle point on the target detection ray that is closest to the target object. Figure 5 (b) Point O is the location of the target object. The third position distance between point A and point O on the target detection ray is obtained.

[0056] In one specific implementation, for each target detection ray, a fourth positional distance is obtained between a first drivable area point and a corresponding boundary point on the target detection ray to obtain a fourth positional difference; wherein there are no obstacle points between the first drivable area point and the corresponding boundary point. For example, such as Figure 6 As shown, Figure 6This is a schematic diagram of an embodiment of the fourth positional difference scenario provided in this application. Figure 6 Point Dk in the diagram represents the corresponding boundary point, which is the corner point between the target detection ray and the preset planar region. Figure 6 Point M is the drivable area point furthest from the target object. The distance between point M and point Dk on the target detection ray is obtained as the fourth position distance.

[0057] Step S13: Using the data point supplementation strategy based on the corresponding data point distribution, supplement the data points on each target detection ray.

[0058] In this embodiment, a data point supplementation strategy based on the distribution of corresponding data points is used to supplement data points on each target detection ray.

[0059] In one embodiment, for each target detection ray, data points are supplemented using different or the same data point supplementation strategy in at least one interval region. The at least one interval region includes at least one of the following: a first interval region between two adjacent drivable area points with a difference threshold greater than a first difference threshold; a second interval region between an obstacle point with a difference threshold greater than a second difference threshold and an adjacent drivable area point; a third interval region between the nearest data point with a difference threshold greater than a third difference threshold and the target object; and a fourth interval region between a first drivable area point with a difference threshold greater than a fourth difference threshold and its corresponding boundary point. The magnitudes of the first, second, third, and fourth difference thresholds are not limited and can be specifically set according to actual usage needs.

[0060] Regarding the first positional distance between each second drivable area point and its adjacent drivable area point included in the first positional difference scenario, if the first positional distance is greater than the first difference threshold, it indicates that the distance between the two second drivable area points corresponding to that first positional distance is too large, determining that there is a missing information between them, and data points need to be added to the first interval area between them on the target detection ray. For example, such as Figure 3 As shown, taking point A as the second drivable area point on the target detection ray and point B as its corresponding adjacent drivable area point as an example: Since the first positional distance between point A and point B is greater than the first difference threshold, data points need to be supplemented in the first interval area between point A and point B on the target detection ray.

[0061] Regarding the second positional difference scenario, which includes the second positional distance between the nearest obstacle point to the target object and its preceding and adjacent drivable area point, if the second positional distance is greater than the second difference threshold, it indicates that the distance between the nearest obstacle point to the target object and its preceding and adjacent drivable area point is too large, confirming a missing information gap. Therefore, data points need to be added to the second gap area between them. For example, such as... Figure 4 As shown, taking point B as the obstacle point closest to the target object on the target detection ray and point A as the drivable area point before and adjacent to it: Since the second position distance between point A and point B is greater than the second difference threshold, data points need to be supplemented in the second interval area between point A and point B on the target detection ray.

[0062] Regarding the third positional difference scenario, specifically the third positional distance between the nearest data point and the target object, if this distance exceeds the third difference threshold, it indicates that the detection blind zone of the detection device is too large. Therefore, data points need to be added to the third interval region between the nearest data point and the target object on the target detection ray. For example, such as... Figure 5 As shown in (a), taking point A as the data point (driving area point) closest to the target object on the target detection ray and point O as the position point of the target object as an example: Since the third position distance between point O and point A is greater than the third difference threshold, the blind zone of the detection device is large. Therefore, it is necessary to supplement the data points in the third interval area between point O and point A on the target detection ray.

[0063] Regarding the fourth positional difference scenario, specifically the fourth positional distance between the farthest drivable area point from the target and its corresponding boundary point, a distance greater than the fourth difference threshold indicates an excessive gap between them. This signifies information loss and necessitates data point supplementation along the target detection ray for the corresponding fourth interval region. For example, if... Figure 6 As shown, taking point M as the drivable area point farthest from the target object on the target detection ray and Dk as the corresponding boundary point: Since the fourth position distance between point M and point Dk is greater than the fourth difference threshold, data points need to be supplemented in the fourth interval area between point M and point Dk on the target detection ray.

[0064] In one specific implementation, the data point supplementation strategy corresponding to the distribution of each data point is the same. In this case, in at least one interval area, different or the same data point supplementation strategy is used to supplement data points. Specifically, in at least one interval area, a number of supplementary drivable area points are added; wherein, the number of supplementary drivable area points are set at equal intervals.

[0065] For example, such as Figure 3 As shown, taking point A as the second drivable area point on the target detection ray and point B as its corresponding adjacent drivable area point as an example: Since the first positional distance between point A and point B is greater than the first difference threshold, data points need to be supplemented in the first interval area between point A and point B on the target detection ray; supplementary drivable area points C1, C2, ..., Cn are added between point A and point B on the target detection ray.

[0066] For example, such as Figure 4 As shown, taking point B as the obstacle point closest to the target object on the target detection ray and point A as the drivable area point before and adjacent to it: Since the second position distance between point A and point B is greater than the second difference threshold, data points need to be added to the second interval area between point A and point B on the target detection ray; supplementary drivable area points C1, C2, ..., Cn are added between point A and point B on the target detection ray.

[0067] For example, such as Figure 5 As shown in (a), taking point A as the data point (drivable area point) closest to the target object on the target detection ray and point O as the position point of the target object as an example: Since the third position distance between point O and point A is greater than the third difference threshold, the blind zone of the detection device is large. Therefore, it is necessary to supplement the data points in the third interval area between point O and point A on the target detection ray; supplementary drivable area points C1, C2, ..., Cn are added between point O and point A on the target detection ray.

[0068] For example, such as Figure 6 As shown, taking point M as the farthest drivable area point on the target detection ray from the target object and point Dk as the corresponding boundary point: Since the fourth position distance between point M and point Dk is greater than the fourth difference threshold, data points need to be added to the fourth interval area between point M and point Dk on the target detection ray; supplementary drivable area points C1, C2, ..., Cn are added between point M and point Dk on the target detection ray.

[0069] Since areas following the nearest obstacle on the target detection ray are necessarily not drivable areas for the target, in one specific implementation, when the data point distribution includes a second positional difference, after supplementing data points on each target detection ray using a data point supplementation strategy corresponding to the data point distribution, data points following the nearest obstacle on the target detection ray will be removed. For example, such as... Figure 4As shown, taking point B as the obstacle point closest to the target object on the target detection ray, point A as the drivable area point before and adjacent to it, and point D as the drivable area point after the nearest obstacle point on the target detection ray: after adding supplementary drivable area points C1, C2, ..., Cn between point A and point B, point D on the target detection ray is removed.

[0070] To improve the accuracy and reliability of the data points supplemented on the target detection rays, in one specific embodiment, data points are supplemented only for target detection rays exhibiting fourth positional differences. Specifically, the angle between each target detection ray and the normal of the detection device is obtained, and detection rays with angles within the specified range are used as reference detection rays. For each reference detection ray, data points are supplemented on each reference detection ray using a data point supplementation strategy corresponding to the fourth positional differences. The size of the angle range is not limited and can be set according to actual usage needs.

[0071] For example, such as Figure 2 As shown, taking the target detection ray as a point ray laid on a preset plane area as an example: obtain the angle between each point ray and the normal of the detection device to obtain the reference point ray [ang-thr, ang+thr] that needs to be supplemented with data points corresponding to the fourth position difference; obtain the intersection of each reference point ray with the preset plane area as the corresponding boundary points D1, D2, ..., Dm-1, Dm, Dm+1, ..., Dn of each reference point ray; for each reference point ray, obtain the fourth position distance between the drivable area point farthest from the target object and the corresponding boundary point on the reference point ray; when there are no obstacle points between the drivable area point farthest from the target object and the corresponding boundary point, and the fourth position distance is greater than the fourth difference threshold, supplement data points in the corresponding fourth interval area (the area between the drivable area point farthest from the target object and the corresponding boundary point) of the reference point ray.

[0072] Step S14: Based on the data points on the target detection ray after each data supplement, determine the drivable area in the target region.

[0073] In this embodiment, the drivable area within the target region is determined based on the data points on the target detection rays after data supplementation. Data point compensation based on the distribution of data points on each target detection ray makes the point cloud data more regular and complete, resolving issues such as data spikes and irregularities caused by detection defects in the detection device. Therefore, the drivable area determined based on the data points on the target detection rays after data supplementation is more accurate and better matches reality.

[0074] In the above embodiments, data point compensation is performed based on the distribution of data points on each target detection ray, making the point cloud data more regular and complete, thus solving the problems of data spikes and irregularities caused by the detection defects of the detection device. Therefore, the drivable area determined based on the data points on the target detection ray after data supplementation is more accurate and more consistent with reality.

[0075] Please see Figure 7 , Figure 7 yes Figure 1 The flowchart shown in step S12 is a schematic diagram of one embodiment. It should be noted that if there are substantially the same result, this embodiment does not necessarily follow the same pattern. Figure 7 The illustrated process sequence is limited. For example... Figure 7 As shown in the embodiments of this application, the target point cloud data is mapped by setting a preset planar region and laying several point rays on the preset planar region, specifically including:

[0076] Step S71: Map the target point cloud data onto a preset planar area to obtain several data points on the preset planar area.

[0077] In this embodiment, the target point cloud data is mapped onto a preset planar area to obtain a number of data points on the preset planar area; wherein, the orthographic projection of the preset planar area on the ground is less than or equal to the orthographic projection of the target area on the ground, and the orthographic projection of the detection device on the preset planar area is within the preset planar area.

[0078] It should be noted that preset planar region parameters (such as preset planar region size, preset planar region resolution, etc.) and the resolution of the point rays laid on the preset planar region (the resolution of a single point ray corresponds to the angular resolution of a target detection ray) can be set according to the actual application scenario requirements. For example, Figure 2 As shown, the grid resolution of the preset planar area is ratio = 0.25m, the grid width is width = 200, and the length is length = 280. Therefore, the preset planar area actually represents an area 70 meters long and 25 meters wide in front of the target object.

[0079] In one embodiment, the target point cloud data is directly projected onto a preset planar area to obtain several data points on the preset planar area; that is, the horizontal and vertical coordinates of the data points on the preset planar area are the same as the horizontal and vertical coordinates of the corresponding 3D points. Of course, in other embodiments, the target point cloud data can also be mapped onto the preset planar area using the following transformation relationship, as shown below:

[0080] grid x =length - [x / ratio]

[0081] gridy =width / 2-y / ratio

[0082] Among them, grid x Represents the x-coordinate on a preset planar region; grid y The x-coordinate represents the vertical coordinate of the point in the preset planar region; the x-coordinate represents the horizontal coordinate of the point in the three-dimensional space; the y-coordinate represents the vertical coordinate of the point in the three-dimensional space; the length represents the length of the preset planar region; the width represents the width of the preset planar region; and the ratio represents the grid resolution of the preset planar region.

[0083] Step S72: For each data point, based on the mapping angle of the data point, map the data point onto the target detection ray that matches the mapping angle of the data point.

[0084] In this embodiment, for each data point, based on the mapping angle of the data point, the data point is mapped to a target detection ray that matches the mapping angle of the data point. In one embodiment, each data point is mapped to a point ray laid out on a preset planar area. The mapping angle of the data point is determined based on the position of the data point on the preset planar area using the following formula:

[0085]

[0086] Among them, angle idx Indicates the mapped angle; angle res The x-axis represents the angular resolution of the target detection ray of the detection device; x represents the x-coordinate of the three-dimensional point in three-dimensional space; y represents the y-coordinate of the three-dimensional point in three-dimensional space.

[0087] For obstacle point cloud data with significant height, it does not actually affect the operation of the target object, so it does not need to be mapped onto the target detection ray, improving the efficiency of determining drivable areas. For example, taking obstacle point cloud data that includes vegetation as an example, in real-world scenarios, tall vegetation exceeding a certain height along the roadside has no actual impact on vehicle movement. Therefore, in one embodiment, the target point cloud data includes obstacle point cloud data. Before mapping each data point onto the target detection ray that matches its mapping angle, data points with heights greater than a height threshold are removed from the non-ground point cloud data. Through this setting, point cloud data that does not affect the actual operation of the target object can be removed, reducing the amount of data that needs to be mapped onto the target detection ray and improving the efficiency of determining drivable areas.

[0088] There is no fixed limit to the height threshold; it can be set according to actual usage needs. It should be noted that if the obstacle point cloud data includes multiple types such as fences and vegetation, the height threshold can be set separately for each type of obstacle point cloud data.

[0089] Please see Figure 8 , Figure 8 yes Figure 1 The diagram shows a flowchart of one embodiment of step S14. It should be noted that if substantially the same result is achieved, the embodiments of this application do not necessarily differ. Figure 8 The illustrated process sequence is limited. For example... Figure 8 As shown, this embodiment includes:

[0090] Step S81: Map the supplementary data points to the preset planar area.

[0091] In this embodiment, the supplementary data points are mapped to a preset planar area.

[0092] Step S82: Determine the outline points of several drivable areas from the data points on the preset planar area.

[0093] In this embodiment, several outline points of drivable areas are determined from data points on a preset planar area. Specifically, each grid point in the preset planar area is traversed sequentially. For a drivable area point, the category of its surrounding ring is determined. If there are obstacle points in its surrounding ring, the drivable area point is determined as an outline point.

[0094] Even after the aforementioned data point supplementation, "holes" will still exist in the preset planar region. Therefore, in one embodiment, a dilation-erosion operation is used to eliminate these "holes" in the preset planar region. Specifically, a suitable convolution kernel is selected and convolved with the preset planar region to obtain a new preset planar region. Subsequently, the drivable area is determined based on the new preset planar region.

[0095] Step S83: Based on the contour points of several drivable areas, obtain several connected regions.

[0096] In this embodiment, several connected components are obtained based on the contour points of several drivable areas. After traversing all grid points in the preset planar area, each connected component is selected from all the contour points. Specifically, the contour points are compared with the centers of the connected components in turn. If one contour point is an eight-neighbor point of another contour point, then the contour point belongs to the current connected component. Otherwise, if the contour point is not adjacent to any connected component point, then the contour point belongs to a new connected component.

[0097] Step S84: From several connected components, remove those whose number of contour points does not meet the requirement, and use the remaining connected components as the drivable area in the target area.

[0098] In this embodiment, connected components whose contour point data do not meet the quantity requirements are removed from a number of connected components, and the remaining connected components are taken as the drivable area in the target region. In other words, connected components whose set of points does not meet the requirements are deleted, and the remaining set of points is the final drivable area contour.

[0099] There is no limit to the quantity requirement; it can be set according to the specific needs of implementation. For example, the quantity requirement is that the number of contour points is less than a quantity threshold, which can be 1, 2, etc.

[0100] Please see Figure 9 , Figure 9 This is a schematic diagram of an embodiment of the drivable area determination device provided in this application. The drivable area determination device 90 includes an acquisition module 91, a determination module 92, a supplementation module 93, and a generation module 94. The acquisition module 91 is used to acquire target point cloud data of the target area; wherein, the target point cloud data is obtained by the detection device using the target detection rays it emits to detect the target area; the determination module 92 is used to map the target point cloud data onto each target detection ray of the detection device, and determine the distribution of data points on each target detection ray based on the data points mapped on each target detection ray; the supplementation module 93 is used to supplement data points on each target detection ray using a data point supplementation strategy corresponding to the data point distribution; the generation module 94 is used to determine the drivable area in the target area based on the data points on each data-supplemented target detection ray.

[0101] The data points include obstacle points and drivable area points; the distribution of data points includes at least one of the following: the first positional difference between two adjacent drivable area points on the target detection ray, the second positional difference between an obstacle point and an adjacent drivable area point on the target detection ray, the third positional difference between the nearest data point on the target detection ray and the target object, and the fourth positional difference between the first drivable area point on the target detection ray and the corresponding boundary point; wherein the detection device is installed on the target object, the nearest data point is the data point closest to the target object, the first drivable area point is the drivable area point farthest from the target object, and the corresponding boundary point is the boundary point of the target area corresponding to the target detection ray.

[0102] The determining module 92 is used to determine the distribution of data points on each target detection ray based on the data points mapped on each target detection ray. Specifically, this includes: for each target detection ray, obtaining the first positional distance between each second drivable area point and its adjacent drivable area point on the target detection ray to obtain a first positional difference; wherein the second drivable area point is located before the nearest obstacle point on the target detection ray, and the corresponding adjacent drivable area point is located after the second drivable area point; and / or, for each target detection ray, obtaining the nearest obstacle point and its adjacent drivable area point on the target detection ray. The second positional spacing between points is used to obtain a second positional difference; wherein the nearest obstacle point is the obstacle point closest to the target object, and the adjacent drivable area point corresponding to the nearest obstacle point is located in front of the nearest obstacle point; and / or, for each target detection ray, the third positional spacing between the nearest data point on the target detection ray and the target object is obtained to obtain a third positional difference; and / or, for each target detection ray, the fourth positional spacing between the first drivable area point on the target detection ray and the corresponding boundary point is obtained to obtain a fourth positional difference; wherein there are no obstacle points between the first drivable area point and the corresponding boundary point.

[0103] The supplementation module 93 is used to supplement data points on each target detection ray by utilizing a data point supplementation strategy based on the distribution of corresponding data points. Specifically, for each target detection ray, data points are supplemented using different or the same data point supplementation strategy in at least one interval region. The at least one interval region includes at least one of the following: a first interval region between two adjacent drivable area points with a difference greater than a first difference threshold; a second interval region between an obstacle point with a difference greater than a second difference threshold and an adjacent drivable area point; a third interval region between the nearest data point with a difference greater than a third difference threshold and the target object; and a fourth interval region between a first drivable area point with a difference greater than a fourth difference threshold and the corresponding boundary point.

[0104] The data point supplementation strategies corresponding to the above data point distributions are the same; the supplementation module 93 is used to supplement data points in at least one interval area using different or the same data point supplementation strategies, specifically including: adding several supplementary drivable area points in at least one interval area; wherein the several supplementary drivable area points are set at equal intervals.

[0105] The data point distribution includes the second position difference situation; the drivable area determination device 90 also includes a rejection module 95, which is used to supplement data points on each target detection ray after using the data point supplementation strategy based on the corresponding data point distribution, specifically including: rejecting data points on the target detection ray that are located after the nearest obstacle point.

[0106] The data point distribution includes the fourth position difference situation; the supplementary module 93 is used to supplement data points on each target detection ray using the data point supplementary strategy corresponding to the data point distribution situation, specifically including: obtaining the angle between each target detection ray and the normal of the detection device, and using the detection rays with the angle within the range of the angle as reference detection rays; for each reference detection ray, supplementing data points on each reference detection ray using the data point supplementary strategy corresponding to the fourth position difference situation.

[0107] The determining module 92 is used to map the target point cloud data onto the target detection rays of the detection device, specifically including: mapping the target point cloud data onto a preset plane area to obtain a number of data points on the preset plane area; wherein the orthographic projection of the preset plane area on the ground is less than or equal to the orthographic projection of the target area on the ground, and the orthographic projection of the detection device on the preset plane area is within the preset plane area; for each data point, based on the mapping angle of the data point, the data point is mapped onto the target detection ray that matches the mapping angle of the data point.

[0108] The aforementioned target point cloud data includes obstacle point cloud data; the elimination module 95 is used to, before mapping each data point to a target detection ray that matches the mapping angle of the data point, specifically include: eliminating data points with heights greater than a height threshold from the obstacle point cloud data.

[0109] The generation module 94 is used to determine the drivable area in the target region based on the data points on the target detection ray after each data supplementation. Specifically, it includes: mapping the supplemented data points to a preset plane region; determining the contour points of several drivable areas from the data points on the preset plane region; obtaining several connected regions based on the contour points of several drivable areas; wherein each connected region includes at least one contour point; and removing connected regions from the several connected regions whose number of contour points does not meet the quantity requirement, and taking the remaining connected regions as the drivable areas in the target region.

[0110] The target point cloud data includes several three-dimensional points and the category of each three-dimensional point; the acquisition module 91 is used to acquire the target point cloud data of the target area, specifically including: acquiring several three-dimensional points collected by the detection device; performing category detection on each three-dimensional point to obtain the category of each three-dimensional point; wherein, the category of the three-dimensional point includes at least the ground category and the obstacle category.

[0111] Please see Figure 10 , Figure 10This is a schematic diagram of an embodiment of the electronic device provided in this application. The electronic device 100 includes a memory 101 and a processor 102 coupled to each other. The processor 102 is used to execute program instructions stored in the memory 101 to implement the steps of the method embodiment for determining any of the drivable areas described above. In a specific implementation scenario, the electronic device 100 may include, but is not limited to, a microcomputer or a server. In addition, the electronic device 100 may also include mobile devices such as laptops and tablets, which are not limited here.

[0112] Specifically, processor 102 controls itself and memory 101 to implement the steps of any of the above-described embodiments of the method for determining a drivable region. Processor 102 may also be referred to as a CPU (Central Processing Unit). Processor 102 may be an integrated circuit chip with signal processing capabilities. Processor 102 may also be a general-purpose processor, digital signal processor (DSP), application-specific integrated circuit (ASIC), field-programmable gate array (FPGA), or other programmable logic device, discrete gate or transistor logic device, or discrete hardware component. A general-purpose processor may be a microprocessor or any conventional processor. Furthermore, processor 102 may be implemented using integrated circuit chips.

[0113] Please see Figure 11 , Figure 11 This is a schematic diagram of the structure of an embodiment of the computer-readable storage medium provided in this application. The computer-readable storage medium 110 of this application embodiment stores program instructions 111. When executed, these program instructions 111 implement the methods provided by any embodiment of the method for determining the drivable area of ​​this application and any non-conflicting combination thereof. The program instructions 111 can form a program file and be stored in the aforementioned computer-readable storage medium 110 in the form of a software product, so that a computer device (which may be a personal computer, server, or network device, etc.) can execute all or part of the steps of the methods of various embodiments of this application. The aforementioned computer-readable storage medium 110 includes various media capable of storing program code, such as a USB flash drive, mobile hard drive, read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disk, or terminal devices such as computers, servers, mobile phones, and tablets.

[0114] If the technical solution of this application involves personal information, the product using this technical solution has clearly informed the user of the personal information processing rules and obtained the user's voluntary consent before processing the personal information. If the technical solution of this application involves sensitive personal information, the product using this technical solution has obtained the user's separate consent before processing the sensitive personal information, and also meets the requirement of "express consent". For example, at personal information collection devices such as cameras, clear and prominent signs are set up to inform users that they have entered the scope of personal information collection and that personal information will be collected. If an individual voluntarily enters the collection scope, it is deemed that they have agreed to the collection of their personal information; or on the personal information processing device, with clear signs / information informing users of the personal information processing rules, authorization is obtained from the individual through pop-up information or by asking the individual to upload their personal information; wherein, the personal information processing rules may include information such as the personal information processor, the purpose of personal information processing, the processing method, and the types of personal information processed.

[0115] The above description is merely an embodiment of this application and does not limit the patent scope of this application. Any equivalent structural or procedural transformations made using the content of this application's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of this application.

Claims

1. A method of determining a drivable area, characterized by, The method comprises: acquiring target point cloud data of a target region; wherein the target point cloud data is obtained by a detection device detecting the target region using target detection rays emitted by the detection device; mapping the target point cloud data to each target detection ray of the detection device, and determining a data point distribution on each target detection ray based on each data point mapped on the target detection ray; wherein the data points include obstacle points and drivable region points, and the data point distribution includes at least one of a first position difference between two adjacent drivable region points on the target detection ray, a second position difference between an obstacle point and an adjacent drivable region point on the target detection ray, a third position difference between a nearest data point on the target detection ray and a target object, and a fourth position difference between a first drivable region point on the target detection ray and a corresponding boundary point; wherein the detection device is arranged on the target object, the nearest data point is the data point closest to the target object, the first drivable region point is the drivable region point farthest from the target object, and the corresponding boundary point is a boundary point of the target region corresponding to the target detection ray; using a data point supplement strategy corresponding to the data point distribution to supplement data points on each target detection ray; determining a drivable region in the target region based on the data points on each target detection ray after data supplement.

2. The method of claim 1, wherein, The determination of the data point distribution on each target detection ray based on each data point mapped on the target detection ray comprises: for each target detection ray, acquiring a first position interval between each second drivable region point and an adjacent drivable region point on the target detection ray to obtain the first position difference; wherein the second drivable region point is located before the nearest obstacle point to the target object on the target detection ray, and the corresponding adjacent drivable region point is located after the second drivable region point; and / or, for each target detection ray, acquiring a second position interval between the nearest obstacle point and an adjacent drivable region point on the target detection ray to obtain the second position difference; wherein the nearest obstacle point is the obstacle point closest to the target object, and the corresponding adjacent drivable region point is located before the nearest obstacle point; and / or, for each target detection ray, acquiring a third position interval between the nearest data point and the target object on the target detection ray to obtain the third position difference; and / or, for each target detection ray, acquiring a fourth position interval between the first drivable region point and the corresponding boundary point on the target detection ray to obtain the fourth position difference; wherein there is no obstacle point between the first drivable region point and the corresponding boundary point.

3. The method according to claim 1 or 2, characterized in that, The data point supplement strategy corresponding to the data point distribution situation is used to supplement data points on each target detection ray, and the method comprises the following steps of: For each target detection ray, data points are supplemented in at least one interval region by using different or same data point supplement strategies, wherein the at least one interval region comprises at least one of the following: a first interval region between the two adjacent drivable region points greater than a first difference threshold, a second interval region between the adjacent drivable region point and the obstacle point greater than a second difference threshold, a third interval region between the nearest data point and the target object greater than a third difference threshold, and a fourth interval region between the first drivable region point and the corresponding boundary point greater than a fourth difference threshold.

4. The method of claim 3, wherein, The data point supplement strategies corresponding to the data point distribution situations are same; and the data points are supplemented in the at least one interval region by using different or same data point supplement strategies, and the method comprises the following steps of: In the at least one interval region, a plurality of supplementary drivable region points are added; wherein the plurality of supplementary drivable region points are arranged at equal intervals.

5. The method of claim 3, wherein, The data point distribution situation comprises a second position difference situation; after the data points are supplemented on each target detection ray by using the data point supplement strategy corresponding to the data point distribution situation, the method further comprises the following steps of: The data points located behind the nearest obstacle point on the target detection ray are removed.

6. The method of claim 1, wherein, The data point distribution situation comprises the fourth position difference situation; the data points are supplemented on each target detection ray by using the data point supplement strategy corresponding to the data point distribution situation, and the method comprises the following steps of: An included angle between each target detection ray and a normal line of the detection device is obtained, and the detection rays with the included angle within an included angle range are taken as reference detection rays; For each reference detection ray, the data points are supplemented on each reference detection ray by using the data point supplement strategy corresponding to the fourth position difference situation.

7. The method of claim 1, wherein, The target point cloud data is mapped onto each target detection ray of the detection device, and the method comprises the following steps of: The target point cloud data is mapped onto a preset planar region to obtain a plurality of data points on the preset planar region; wherein a normal projection of the preset planar region on the ground is less than or equal to a normal projection of the target region on the ground, and a normal projection of the detection device in the preset planar region is in the preset planar region; For each data point, the data point is mapped onto a target detection ray matched with a mapping angle of the data point based on the mapping angle of the data point.

8. The method of claim 7, wherein, The target point cloud data comprises obstacle point cloud data; before the data point is mapped onto the target detection ray matched with the mapping angle of the data point based on the mapping angle of the data point, the method further comprises the following steps of: From the obstacle point cloud data, data points with a height greater than a height threshold are removed.

9. The method of claim 7, wherein, The drivable region in the target region is determined based on the data points on each data-supplemented target detection ray, and the method comprises the following steps of: mapping the supplementary data points to the preset planar region; determining contour points of a plurality of drivable regions from the data points on the preset planar region; obtaining a plurality of connected domains based on the contour points of the plurality of drivable regions; wherein the connected domain includes at least one contour point; eliminating the connected domain including the contour points not meeting the quantity requirement from the plurality of connected domains, and taking the remaining connected domain as the drivable region in the target region.

10. The method of claim 1, wherein, The target point cloud data includes a plurality of three-dimensional points and the category of each three-dimensional point; the target point cloud data of the target region includes: obtaining a plurality of three-dimensional points collected by the detection device; detecting the category of each three-dimensional point to obtain the category of each three-dimensional point; wherein the category of the three-dimensional point includes at least ground and obstacle.

11. A drivable region determination device, characterized in that, an acquisition module is configured to acquire target point cloud data of a target region; wherein the target point cloud data is obtained by a detection device detecting the target region using target detection rays emitted by the detection device; a determination module is configured to map the target point cloud data to each target detection ray of the detection device, and determine the data point distribution on each target detection ray based on the data points mapped on each target detection ray; wherein the data points include obstacle points and drivable region points, and the data point distribution includes at least one of the following: a first position difference between two adjacent drivable region points on the target detection ray, a second position difference between an obstacle point and an adjacent drivable region point on the target detection ray, a third position difference between the nearest data point and a target object on the target detection ray, and a fourth position difference between a first drivable region point and a corresponding boundary point on the target detection ray; wherein the detection device is arranged on the target object, the nearest data point is the data point closest to the target object, the first drivable region point is the drivable region point farthest from the target object, and the corresponding boundary point is the boundary point of the target region corresponding to the target detection ray; a supplement module is configured to supplement data points on each target detection ray using a data point supplement strategy corresponding to the data point distribution; a generation module is configured to determine a drivable region in the target region based on the data points on each data-supplemented target detection ray.

12. An electronic device, comprising: The electronic device includes a processor and a memory, the memory stores program instructions, and the processor is configured to execute the program instructions to implement the drivable region determination method of any one of claims 1-10.

13. A computer-readable storage medium, characterized in that, The computer readable storage medium is used to store program instructions, and the program instructions can be executed to implement the drivable region determination method of any one of claims 1-10.

Images