Map generation method, device, medium and electronic equipment

Abstract

The present disclosure relates to a map generation method, device, medium and electronic device, the method comprising: obtaining a multi-frame point cloud collected by a point cloud acquisition device in a target area; determining whether the signal strength of the target area meets the preset high-intensity requirement; determining a point cloud matching pair from the multi-frame point cloud according to the result of determining whether the signal strength of the target area meets the preset high-intensity requirement, and determining the target pose information of each of the multi-frame point clouds according to the point cloud matching pair, wherein the same points exist between the two frames of point clouds constituting the point cloud matching pair; generating a map of the target area according to the target pose information of each of the multi-frame point clouds. Through the above technical scheme, by constructing a point cloud matching pair, determining the target pose information of each of the multi-frame point clouds according to the point cloud matching pair, the dependence on the global positioning system is reduced, and there is no need to find the nearest point for each point to match as in the related art. The amount of calculation is relatively small, which can improve the efficiency and accuracy of map generation.

Description

Technical Field

[0001] The present disclosure relates to the field of autonomous driving, and in particular, to a map generation method, device, medium and electronic device. Background Art

[0002] Maps have a wide range of application scenarios. For example, driverless vehicles can use electronic maps to recognize the surrounding environment, thereby controlling the steering, acceleration or deceleration of driverless vehicles. With the rise of driverless vehicles, the scale of map collection is expanding. Taking the takeaway scenario as an example, driverless vehicles can perform delivery tasks in the city and deliver items to users, which can achieve contactless delivery. Since driverless vehicles need maps to recognize the surrounding environment, the reliable operation of driverless vehicles depends on high-precision maps. However, due to the diversity of urban scenarios, it is difficult to ensure the accuracy and quality of the constructed maps. Summary of the invention

[0003] The purpose of the present disclosure is to provide a map generation method, device, medium and electronic device to partially solve the above-mentioned problems existing in the related art.

[0004] In order to achieve the above objectives, in a first aspect, the present disclosure provides a map generation method, the method comprising:

[0005] Acquire multiple frames of point clouds collected by a point cloud collection device in a target area;

[0006] Determining whether the signal strength of the target area meets a preset high strength requirement;

[0007] Determining a point cloud matching pair from the multiple point clouds according to a result of determining whether the signal strength of the target area meets a preset high-intensity requirement, and determining target pose information of each of the multiple point clouds according to the point cloud matching pair, wherein there are identical points between two point clouds constituting the point cloud matching pair;

[0008] A map of the target area is generated according to the target pose information of each of the multiple frame point clouds.

[0009] Optionally, when the determination result indicates that the signal strength of the target area meets a preset high strength requirement, determining a point cloud matching pair from the multi-frame point cloud includes:

[0010] For each frame point cloud, other frame point clouds in the multiple frame point clouds that have the same points as the frame point cloud and have the farthest timestamp distance from the frame point cloud form the point cloud matching pair with the frame point cloud.

[0011] Optionally, when the determination result indicates that the signal strength of the target area meets a preset high-intensity requirement, determining the target pose information of each of the multiple point cloud frames according to the point cloud matching pair includes:

[0012] For each of the point cloud matching pairs, determine the spliced ​​point clouds of the two frames of point clouds constituting the point cloud matching pair, and determine the relative pose information between the two frames of point clouds based on the spliced ​​point clouds of the two frames of point clouds, wherein the spliced ​​point cloud of each frame of point cloud is formed by splicing point clouds collected by the point cloud acquisition device in a specified area, and the specified area is an area around the position where the point cloud acquisition device collects the frame of point cloud;

[0013] The target pose information of each of the multiple frame point clouds is determined according to the relative pose information corresponding to each point cloud matching pair.

[0014] Optionally, when the determination result indicates that the signal strength of the target area does not meet a preset high strength requirement, determining a point cloud matching pair from the multi-frame point cloud includes:

[0015] For each frame point cloud, other frame point clouds in the multiple frame point clouds that have the same points as the frame point cloud are combined with the frame point cloud to form the point cloud matching pair.

[0016] Optionally, when the determination result indicates that the signal strength of the target area does not meet the preset high strength requirement, determining the target pose information of each of the multiple frame point clouds according to the point cloud matching pair includes:

[0017] For each of the point cloud matching pairs, the feature points of the two frames of point clouds constituting the point cloud matching pair are determined, and the same-name points in the two frames of point clouds are determined from the feature points; with the goal of minimizing the distance information between the same-name points, the current posture information of the two frames of point clouds is iteratively adjusted until the distance information between the same-name points is less than a preset distance threshold, thereby obtaining the target posture information of the two frames of point clouds.

[0018] Optionally, determining whether the signal strength of the target area meets a preset high strength requirement includes:

[0019] Determine first device pose information at first preset time intervals during the process of the point cloud acquisition device acquiring the multiple frames of point clouds in the target area, and first confidences corresponding to each piece of the first device pose information;

[0020] It is determined, according to the first confidence level, whether the signal strength of the target area meets a preset high strength requirement.

[0021] Optionally, the determining of first device pose information at intervals of a first preset time during the process of the point cloud acquisition device acquiring the multiple frames of point clouds in the target area, and first confidences corresponding to each piece of the first device pose information, respectively, includes:

[0022] Acquire second device position information of the navigation device collected by the navigation device in the target area at intervals of a second preset time, and second confidences corresponding to each piece of the second device position information;

[0023] Determine the first device posture information according to the relative position relationship between the navigation device and the point cloud acquisition device, and the second device posture information;

[0024] For each first device posture information, determine the target second device posture information in the second device posture information that is closest to the timestamp of the first device posture information, and use the second confidence of the target second device posture information as the first confidence corresponding to the first device posture information.

[0025] Optionally, determining, according to the first confidence level, whether the signal strength of the target area meets a preset high strength requirement includes:

[0026] If the proportion of the first confidences above the preset confidence threshold is greater than the preset proportion threshold, or the number of the first confidences above the preset confidence threshold is greater than the preset number threshold, it is determined that the signal strength of the target area meets the preset high strength requirement.

[0027] In a second aspect, the present disclosure provides a map generating device, the device comprising:

[0028] An acquisition module is configured to acquire multiple frames of point clouds collected by a point cloud acquisition device in a target area;

[0029] A first determination module is configured to determine whether the signal strength of the target area meets a preset high strength requirement;

[0030] A second determination module is configured to determine a point cloud matching pair from the multiple point clouds according to a determination result of whether the signal strength of the target area meets a preset high strength requirement, and determine target pose information of each of the multiple point clouds according to the point cloud matching pair, wherein there are identical points between two point clouds constituting the point cloud matching pair;

[0031] A generation module is configured to generate a map of the target area based on the target pose information of each of the multiple frame point clouds.

[0032] Optionally, when the determination result indicates that the signal strength of the target area meets a preset high-intensity requirement, the second determination module is used to determine a point cloud matching pair from the multi-frame point cloud in the following manner: for each frame point cloud, other frame point clouds in the multi-frame point cloud that have the same points as the frame point cloud and have the farthest timestamp distance from the frame point cloud constitute the point cloud matching pair with the frame point cloud.

[0033] Optionally, when the determination result indicates that the signal strength of the target area meets a preset high strength requirement, the second determination module includes:

[0034] A first determination submodule is configured to determine, for each point cloud matching pair, respective spliced ​​point clouds of two frames of point clouds constituting the point cloud matching pair, and determine relative pose information between the two frames of point clouds based on the respective spliced ​​point clouds of the two frames of point clouds, wherein the spliced ​​point cloud of each frame of point cloud is spliced ​​by point clouds collected by the point cloud acquisition device in a specified area, and the specified area is an area around a position where the point cloud acquisition device collects the frame of point cloud;

[0035] The second determination submodule is configured to determine the target pose information of each of the multiple frames of point clouds according to the relative pose information corresponding to each point cloud matching pair.

[0036] Optionally, when the determination result indicates that the signal strength of the target area does not meet the preset high intensity requirement, the second determination module is used to determine a point cloud matching pair from the multi-frame point cloud in the following manner: for each frame point cloud, other frame point clouds in the multi-frame point cloud that have the same points as the frame point cloud are combined with the frame point cloud to form the point cloud matching pair.

[0037] Optionally, when the determination result indicates that the signal strength of the target area does not meet a preset high strength requirement, the second determination module includes:

[0038] The third determination submodule is configured to determine, for each of the point cloud matching pairs, the feature points of the two frames of point clouds constituting the point cloud matching pair, and determine the same-name points in the two frames of point clouds from the feature points; with the goal of minimizing the distance information between the same-name points, iteratively adjust the current posture information of the two frames of point clouds until the distance information between the same-name points is less than a preset distance threshold, thereby obtaining the target posture information of the two frames of point clouds.

[0039] Optionally, the first determining module includes:

[0040] A fourth determination submodule is configured to determine first device posture information of each first preset time period during the process of the point cloud acquisition device acquiring the multiple frames of point clouds in the target area, and first confidences corresponding to each of the first device posture information;

[0041] The fifth determination submodule is configured to determine whether the signal strength of the target area meets a preset high strength requirement according to the first confidence level.

[0042] Optionally, the fourth determining submodule includes:

[0043] an acquisition submodule, configured to acquire second device posture information of the navigation device collected by the navigation device in the target area at second preset time intervals, and second confidences corresponding to each piece of the second device posture information;

[0044] a sixth determination submodule, configured to determine the first device posture information according to the relative position relationship between the navigation device and the point cloud acquisition device, and the second device posture information;

[0045] The seventh determination submodule is configured to determine, for each first device posture information, the target second device posture information in the second device posture information that is closest to the timestamp of the first device posture information, and use the second confidence of the target second device posture information as the first confidence corresponding to the first device posture information.

[0046] Optionally, the fifth determination module is used to: if the proportion of the first confidence levels higher than a preset confidence threshold is greater than a preset proportion threshold, or the number of first confidence levels higher than the preset confidence threshold is greater than a preset number threshold, then determine that the signal strength of the target area meets the preset high intensity requirement.

[0047] In a third aspect, the present disclosure provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method provided in the first aspect of the present disclosure.

[0048] In a fourth aspect, the present disclosure provides an electronic device, comprising: a memory on which a computer program is stored; and a processor for executing the computer program in the memory to implement the steps of the method provided in the first aspect of the present disclosure.

[0049] Through the above technical solution, the multi-frame point cloud collected by the point cloud acquisition device in the target area is obtained, and it is determined whether the signal strength of the target area meets the preset high-intensity requirement. According to the determination result, the point cloud matching pair is determined from the multi-frame point cloud, and the target pose information of each of the multi-frame point clouds is determined according to the point cloud matching pair, and the map of the target area is generated according to the target pose information of each of the multi-frame point clouds. In this way, by constructing the point cloud matching pair and determining the target pose information of each of the multi-frame point clouds according to the point cloud matching pair, there is no need to directly obtain the pose information of the point cloud acquisition device according to the pose information of the global positioning system and the inertial measurement unit, which reduces the dependence on the global positioning system, and can generate a map with higher accuracy even when the signal of the global positioning system is weak. In addition, if there are identical points between two frames of point clouds, these two frames of point clouds can construct point cloud matching pairs. There is no need to find the nearest point to match each point as in the related art. The amount of calculation is relatively small. Based on the determination result of whether the signal strength of the target area meets the preset high-intensity requirement, point cloud matching pairs are determined from multiple frames of point clouds. This allows the number of constructed point cloud matching pairs to be adapted to the signal strength of the target area. By constructing point cloud matching pairs, the efficiency and accuracy of map generation can be improved.

[0050] Other features and advantages of the present disclosure will be described in detail in the following detailed description. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] The accompanying drawings are used to provide a further understanding of the present disclosure and constitute a part of the specification. Together with the following specific embodiments, they are used to explain the present disclosure but do not constitute a limitation of the present disclosure. In the accompanying drawings:

[0052] Figure 1 The figure is a flowchart of a map generating method according to an exemplary embodiment.

[0053] Figure 2 The present invention is a flowchart of a map generating method according to another exemplary embodiment.

[0054] Figure 3 The invention is a flowchart of a method for determining first device posture information and a first confidence level according to an exemplary embodiment.

[0055] Figure 4a It is a schematic diagram of a map generated by using an implementation method in the related art.

[0056] Figure 4b It is a schematic diagram of a map generated by using an implementation method in the related art.

[0057] Figure 4c is a schematic diagram of a map generated using an embodiment of the present disclosure.

[0058] Figure 5The figure is a block diagram of a map generating device according to an exemplary embodiment.

[0059] Figure 6 It is a block diagram of an electronic device according to an exemplary embodiment.

[0060] Figure 7 It is a block diagram of an electronic device according to an exemplary embodiment. DETAILED DESCRIPTION

[0061] In the related art, the map construction system generally collects map data through vehicle-mounted laser radar. The position and posture of the vehicle-mounted laser radar is interpolated from the post-processed position and posture of the GPS (Global Position System) and IMU (Inertial Measurement Unit), so it is highly dependent on GPS. The system can provide centimeter-level accuracy in open areas, such as highways, but in scenes such as urban high-rise buildings and tree-lined roads, satellite signals are easily missing or interfered with. Directly using GPS and IMU to obtain the position and posture of the vehicle-mounted laser radar will result in inaccurate radar position and posture in weak GPS scenes, and the constructed map will have ghosting, affecting the accuracy of the map. Moreover, in the related art, when optimizing the radar position and posture, it is necessary to match each point in the point cloud with the point closest to it. On the one hand, this method has a very large amount of calculation and low efficiency. On the other hand, it is easily interfered by the noise of dynamic vehicles, resulting in failure to match the point at the closest distance.

[0062] In view of this, the present disclosure provides a map generation method, device, medium and electronic device to partially solve the above-mentioned problems existing in the related art.

[0063] The specific implementation of the present disclosure is described in detail below in conjunction with the accompanying drawings. It should be understood that the specific implementation described herein is only used to illustrate and explain the present disclosure, and is not used to limit the present disclosure.

[0064] Figure 1 is a flowchart of a map generation method according to an exemplary embodiment, which can be applied to an electronic device with processing capabilities, such as a terminal or a server. Figure 1 As shown, the method may include S101 to S104.

[0065] In S101, multiple frames of point clouds collected by a point cloud collection device in a target area are obtained.

[0066] The target area can be any area where a map needs to be constructed, such as a street in a city, or a road in a park or community, etc. The point cloud acquisition device can be, for example, a vehicle-mounted laser radar. For example, a vehicle integrated with a point cloud acquisition device can travel one or more round trips in the target area. During the driving of the vehicle, the point cloud acquisition device can continuously scan and collect the surrounding scenes. Each scan of the point cloud acquisition device is a frame. As the vehicle drives, the point cloud acquisition device can collect multiple frames of point clouds in the target area. The present disclosure does not limit the number of multiple frames of point clouds.

[0067] In S102, it is determined whether the signal strength of the target area meets a preset high strength requirement.

[0068] For example, whether the signal strength in the target area meets the preset high-intensity requirement can be determined based on the confidence level of the device posture information of the point cloud acquisition device in the target area. The confidence level can represent the credibility of the device posture information. The higher the confidence level, the higher the signal strength in the target area.

[0069] In S103, based on the result of determining whether the signal strength of the target area meets the preset high strength requirement, point cloud matching pairs are determined from the multiple frame point clouds, and based on the point cloud matching pairs, target pose information of each of the multiple frame point clouds is determined.

[0070] There are common points between the two frames of point clouds that constitute a point cloud matching pair. The point cloud acquisition device can scan the same object at different times. For example, if the same tree is scanned at the first time and the second time, then there are common points between the frame of point cloud acquired by the point cloud acquisition device at the first time and the frame of point cloud acquired at the second time, and these two frames of point clouds can constitute a point cloud matching pair (frame pair).

[0071] For example, the present disclosure may detect point cloud matching pairs based on a closed loop. For example, a vehicle integrated with a point cloud acquisition device may travel one or more round trips in a target area. During the vehicle's travel in direction 1 in the target area, the point cloud acquisition device acquires a frame of point cloud at a first moment. During the vehicle's travel in direction 2 in the target area, the point cloud acquisition device acquires a frame of point cloud at a second moment. Direction 1 and direction 2 may be opposite. As in the above example, if the point cloud acquired by the point cloud acquisition device at the first moment and the point cloud acquired at the second moment have the same points, then the two frames of point clouds may constitute a point cloud matching pair, that is, the two frames of point clouds that constitute the point cloud matching pair may not be two frames of point clouds that are adjacent in time. In this way, based on the closed-loop detection of point cloud matching pairs, the detection range is wider when constructing point cloud matching pairs, and is not limited to the constraints of adjacent frames.

[0072] Furthermore, based on the result of determining whether the signal strength of the target area meets the preset high-intensity requirement, a point cloud matching pair is determined from multiple frames of point clouds. When the signal strength of the target area does not meet the preset high-intensity requirement, it can be characterized that the signal strength of the target area is not high enough. In order to avoid the influence of weak signal strength on the map quality, more point cloud matching pairs can be constructed. When the signal strength of the target area meets the preset high-intensity requirement, it can be characterized that the signal strength of the target area is high. In order to improve computing efficiency, the number of point cloud matching pairs constructed in this case can be relatively small.

[0073] There can be multiple point cloud matching pairs. Based on the point cloud matching pairs, the target pose information of each frame of point cloud can be determined. The determination results of whether the signal strength of the target area meets the preset high-intensity requirements are different. The method of determining the target pose information of each frame of point cloud can be different, so that the number of constructed point cloud matching pairs and the method of determining the target pose information can be adapted to the signal strength of the target area. The target pose information of a frame of point cloud can refer to the position information and attitude information of the point cloud acquisition device when the frame of point cloud is collected. The position information can be three-dimensional coordinate information, and the attitude information can include pitch angle, roll angle, and yaw angle.

[0074] In S104, a map of the target area is generated according to the target pose information of each of the multiple frame point clouds.

[0075] After determining the target pose information of each frame of point cloud, the target pose information of each frame of point cloud can be converted to the global coordinate system. If a three-dimensional map needs to be generated, a three-dimensional map of the target area can be generated based on the pose information converted to the global coordinate system. If a two-dimensional map needs to be generated, the pose information in the global coordinate system can be converted to a two-dimensional plane to obtain a two-dimensional map of the target area.

[0076] Through the above technical solution, the multi-frame point cloud collected by the point cloud acquisition device in the target area is obtained, and it is determined whether the signal strength of the target area meets the preset high-intensity requirement. According to the determination result, the point cloud matching pair is determined from the multi-frame point cloud, and the target pose information of each of the multi-frame point clouds is determined according to the point cloud matching pair, and the map of the target area is generated according to the target pose information of each of the multi-frame point clouds. In this way, by constructing the point cloud matching pair and determining the target pose information of each of the multi-frame point clouds according to the point cloud matching pair, there is no need to directly obtain the pose information of the point cloud acquisition device according to the pose information of the global positioning system and the inertial measurement unit, which reduces the dependence on the global positioning system, and can generate a map with higher accuracy even when the signal of the global positioning system is weak. In addition, if there are identical points between two frames of point clouds, these two frames of point clouds can construct point cloud matching pairs. There is no need to find the nearest point to match each point as in the related art. The amount of calculation is relatively small. Based on the determination result of whether the signal strength of the target area meets the preset high-intensity requirement, point cloud matching pairs are determined from multiple frames of point clouds. This allows the number of constructed point cloud matching pairs to be adapted to the signal strength of the target area. By constructing point cloud matching pairs, the efficiency and accuracy of map generation can be improved.

[0077] Figure 2 is a flowchart of a map generation method according to another exemplary embodiment. Figure 2 As shown, the method may include S201 to S209, wherein S102 may include S202 and S203.

[0078] In S201, multiple frames of point clouds collected by a point cloud collection device in a target area are obtained. The implementation of step S201 may refer to S101.

[0079] In S202, first device pose information at intervals of a first preset time period during a process in which a point cloud acquisition device acquires multiple frames of point clouds in a target area, and first confidence levels corresponding to respective pieces of first device pose information are determined.

[0080] The exemplary implementation of step S202 may be as follows: Figure 3 As shown, including S2021 to S2023.

[0081] In S2021, second device position information of the navigation device collected by the navigation device in the target area every second preset time period and second confidence levels corresponding to each piece of second device position information are obtained.

[0082] Among them, the navigation device can be a combined navigation device, for example, it can be a combined navigation device composed of GPS and IMU. The navigation device and the point cloud acquisition device can be integrated on the vehicle at the same time. During the vehicle's driving in the target area, the navigation device can collect its own second device posture information every second preset time period. The second preset time period can be, for example, 0.01s, that is, the navigation device can collect its own posture information at a frequency of 100HZ.

[0083] When the navigation device outputs the second device posture information collected by itself, it can also output the second confidence level corresponding to each second device posture information. The second confidence level can be used to characterize the credibility of the corresponding second device posture information. The higher the second confidence level, the higher the credibility of the corresponding second device posture information, that is, the higher the accuracy. For example, the second confidence level can be determined based on parameters such as the number of satellites, longitude and latitude errors, and position accuracy when the navigation device collects the corresponding second device posture information.

[0084] In S2022, the first device posture information is determined according to the relative position relationship between the navigation device and the point cloud acquisition device, and the second device posture information.

[0085] The relative position relationship between the navigation device and the point cloud acquisition device can be calibrated in advance. According to the relative position relationship and the second device posture information of the navigation device, the first device posture information of the point cloud acquisition device at intervals of the first preset time length can be obtained by interpolation. Among them, the frequency of the navigation device and the frequency of the point cloud acquisition device are different. The frequency of the navigation device can be 100HZ, and the frequency of the point cloud acquisition device can be 10HZ. The first preset time length can be 0.1s, that is, the first device posture information of the low-frequency point cloud acquisition device can be determined based on the second device posture information of the high-frequency navigation device. It should be noted that the above examples of the first preset time length and the second preset time length are for explanation only and do not constitute a limitation to the embodiments of the present disclosure.

[0086] In S2023, for each first device posture information, determine the target second device posture information in the second device posture information that is closest to the timestamp of the first device posture information, and use the second confidence of the target second device posture information as the first confidence corresponding to the first device posture information.

[0087] Among them, the point cloud acquisition device and the navigation device have different frequencies. The second confidence of the target second device posture information closest to the timestamp of the first device posture information can best characterize the signal strength at the corresponding moment of the first device posture information. Therefore, the second confidence of the target second device posture information can be used as the first confidence corresponding to the first device posture information. In this way, the confidence of the posture information of the point cloud acquisition device can be accurately determined according to the confidence of the posture information of the navigation device.

[0088] In S203, according to the first confidence level, it is determined whether the signal strength of the target area meets the preset high strength requirement. If not, S204, S205 and S209 are executed; if yes, S206 to S209 are executed.

[0089] An exemplary implementation of step S203 may be: if the proportion of first confidences above a preset confidence threshold is greater than a preset proportion threshold, or the number of first confidences above the preset confidence threshold is greater than a preset number threshold, it is determined that the signal strength of the target area meets the preset high strength requirement.

[0090] Among them, the first confidence is higher than the preset confidence threshold, which can indicate that the corresponding first device posture information has a higher degree of credibility. If the proportion of the first confidences higher than the preset confidence threshold is greater than the preset proportion threshold, or the number of the first confidences higher than the preset confidence threshold is greater than the preset number threshold, it can be indicated that in the process of collecting point cloud in the target area, the overall credibility of the first device posture information of the point cloud collection device is higher. The first device posture information of the point cloud collection device can be determined based on the second device posture information of the navigation device. Therefore, it can also be indicated that the posture information collected by the navigation device itself is relatively accurate, that is, the signal strength in the target area is better and meets the preset high-intensity requirements.

[0091] In S204, for each frame point cloud, other frame point clouds in multiple frame point clouds that have the same points as the frame point cloud are combined with the frame point cloud to form a point cloud matching pair.

[0092] Among them, the signal strength of the target area does not meet the preset high-intensity requirement, which can indicate that the signal strength of the target area is not high enough. In order to avoid the influence of weak signal strength on the map quality, more point cloud matching pairs can be constructed. When the determination result indicates that the signal strength of the target area does not meet the preset high-intensity requirement, the implementation method of determining the point cloud matching pairs from multiple frames of point clouds in S103 can be the same as S204, that is, as long as there are the same points as the frame point cloud, a point cloud matching pair can be formed with the frame point cloud, so that the number of formed point cloud matching pairs is greater, so as to improve the quality of the finally generated map.

[0093] In S205, for each point cloud matching pair, the feature points of the two frames of point clouds constituting the point cloud matching pair are determined, and the same-name points in the two frames of point clouds are determined from the feature points; with the goal of minimizing the distance information between the same-name points, the current posture information of the two frames of point clouds is iteratively adjusted until the distance information between the same-name points is less than the preset distance threshold, thereby obtaining the target posture information of the two frames of point clouds.

[0094] When the determination result indicates that the signal strength of the target area does not meet the preset high strength requirement, the implementation method of determining the target pose information of each of the multiple frame point clouds according to the point cloud matching pairs in S103 may be as in S205.

[0095] Among them, the feature points can be key points. For example, deep learning (Deep Registration) can be used to extract feature points in the point cloud. The point cloud includes a large number of points. First, the feature points of each frame of the point cloud are determined, and then the points with the same name are determined from the feature points. This method can reduce a certain amount of calculation.

[0096] Among them, the same-name points can refer to the same points. There can be one or more groups of same-name points in two frames of point clouds. The distance information between the same-name points can refer to the Euclidean distance between the same-name points, or the distance between the same-name points in the feature space. For example, vectors can be used to represent the points, and the distance information between the same-name points can be converted into the distance between the vectors corresponding to the two points. Since the same-name points are the same points, the distance between the same points should be as small as possible. Therefore, in the present disclosure, the current posture information of each of the two frames of point clouds is iteratively adjusted with the goal of minimizing the distance information between the same-name points. For example, the iterative adjustment method can adopt the gradient descent method, wherein the initial posture information of the point cloud can be the device posture information of the point cloud acquisition device when the frame of point cloud is collected. If there are multiple groups of same-name points in the two frames of point clouds, the sum of the distance information between each group of same-name points can be minimized as the goal, and the current posture information of each of the two frames of point clouds can be iteratively adjusted. The condition for exiting the iterative adjustment can be that the distance information between the same-name points is less than the preset distance threshold. When the distance information between the same-name points is less than the preset distance threshold, it indicates that the distance between the same-name points meets the requirement, and the iterative adjustment process can be exited to obtain the target pose information of the two frame point clouds. The objective function E in the iterative adjustment process can be shown as the following formula (1):

[0097]

[0098] Among them, m represents the m-th frame point cloud, n represents the n-th frame point cloud, the m-th frame point cloud and the n-th frame point cloud constitute a point cloud matching pair, and FPS represents the set of point cloud matching pairs. represents the ath point in the mth frame point cloud, represents the bth point in the nth frame point cloud. The ath point and the bth point are a group of points with the same name. ab Indicates the weight of the group of points with the same name, Represents the distance information between the a-th point and the b-th point, S k represents the set of points with the same name in the m-th frame point cloud and the n-th frame point cloud, and T represents the transposed matrix.

[0099] In S206, for each frame point cloud, other frame point clouds in multiple frame point clouds that have the same points as the frame point cloud and have the farthest timestamp distance from the frame point cloud are combined with the frame point cloud to form a point cloud matching pair.

[0100] Among them, the signal strength of the target area meets the preset high-intensity requirement, which can be used to characterize that the signal strength of the target area is high. In order to improve the calculation efficiency, the number of point cloud matching pairs constructed in this case can be relatively small. When constructing point cloud matching pairs, in addition to considering whether there are identical points, the acquisition time of the point cloud can also be considered. In the case where the signal strength of the target area characterized by the determination result meets the preset high-intensity requirement, the implementation method of determining point cloud matching pairs from multiple frame point clouds in S103 can be as in S206. For each frame point cloud, other frame point clouds in the multiple frame point clouds that have the same points as the frame point cloud and are farthest from the timestamp of the frame point cloud constitute a point cloud matching pair with the frame point cloud, wherein the timestamp of the frame point cloud can refer to the time when the point cloud acquisition device acquires the frame point cloud, and each frame point cloud has a corresponding timestamp. The number of point cloud matching pairs constructed in this way is relatively small, which can improve the efficiency of map generation.

[0101] In S207, for each point cloud matching pair, the spliced ​​point clouds of the two frames of point clouds constituting the point cloud matching pair are determined, and the relative pose information between the two frames of point clouds is determined based on the spliced ​​point clouds of the two frames of point clouds.

[0102] When the determination result indicates that the signal strength of the target area meets the preset high strength requirement, the implementation method of determining the target pose information of each of the multiple frame point clouds according to the point cloud matching pairs in S103 may be as described in S207 and S208.

[0103] The stitched point cloud of each frame of point cloud is stitched together by the point cloud acquisition device in the specified area, and the specified area can be the area around the position where the point cloud acquisition device is when acquiring the frame of point cloud. For example, the position where the point cloud acquisition device is when acquiring the frame of point cloud, and the area formed by 5m in front of and 5m behind the position can be used as the specified area. It should be noted that this value is only used as an example, and there is no limitation on the scope of the specified area.

[0104] The stitched point cloud is formed by stitching together multiple frames of point clouds. The relative pose information between the two frames of point clouds is determined based on the stitched point clouds of the two frames of point clouds. The scene where each frame of point cloud is located can be expanded, so that the field of view corresponding to each frame of point cloud is wider, thereby making the relative pose information between the two frames of point clouds more accurate. For example, GICP (Generalized Iterative Closest Point) can be used to determine the relative pose information between the two frames of point clouds based on the stitched point cloud.

[0105] In S208, target pose information of each of the multiple frames of point clouds is determined according to the relative pose information corresponding to each point cloud matching pair.

[0106] After obtaining the relative pose information corresponding to each point cloud matching pair, for example, Pose Graph can be used to obtain the target pose information of multiple frame point clouds. The objective function can be expressed as the following formula (2):

[0107]

[0108] Among them, i represents the i-th frame point cloud, j represents the j-th frame point cloud, the i-th frame point cloud and the j-th frame point cloud constitute a point cloud matching pair, x i Represents the pose information of the i-th frame point cloud, x j represents the pose information of the j-th frame point cloud, T i,j represents the rotation and translation matrix of the i-th frame point cloud relative to the j-th frame point cloud, c i,j Represents the difference between the pose information of the i-th frame point cloud after the rotation and translation matrix transformation and the pose information of the j-th frame point cloud.

[0109] In S209, a map of the target area is generated according to the target pose information of each of the multiple frame point clouds. The implementation of step S209 can refer to S104.

[0110] Through the above technical solution, the signal strength of the target area meets the preset high-intensity requirement or does not meet the preset high-intensity requirement. In these two cases, the methods of determining the point cloud matching pairs are different, and the methods of determining the target pose information of each of the multiple-frame point clouds are also different. When the signal strength of the target area meets the preset high-intensity requirement, the number of constructed point cloud matching pairs can be relatively small. When the signal strength of the target area does not meet the preset high-intensity requirement, in order to improve the map accuracy, the number of constructed point cloud matching pairs can be relatively large. In this way, by constructing point cloud matching pairs, the efficiency and accuracy of map generation can be improved.

[0111] The following is a comparison between a map generated by using an implementation in the related art and a map generated by using an implementation of the present disclosure. Figure 4a and 4b is a schematic diagram of a map generated by using an implementation method in the related art, such as Figure 4a and Figure 4b As shown in the figure, it can be seen from the area in the white rectangle that the lane lines in both maps have obvious ghosting. Figure 4c is a schematic diagram of a map generated by using an embodiment of the present disclosure, such as Figure 4c As shown, from Figure 4c The area in the rectangular box on the left shows that the lane line does not have a ghosting phenomenon. Figure 4cAs can be seen from the area in the rectangular box on the middle right, the objects are not blurred and the map quality and accuracy are higher.

[0112] Based on the same inventive concept, the present disclosure also provides a map generating device. Figure 5 is a block diagram of a map generating device according to an exemplary embodiment. Figure 5 As shown, the device 500 may include:

[0113] An acquisition module 501 is configured to acquire multiple frames of point clouds collected by a point cloud acquisition device in a target area;

[0114] A first determination module 502 is configured to determine whether the signal strength of the target area meets a preset high strength requirement;

[0115] The second determination module 503 is configured to determine a point cloud matching pair from the multiple point clouds according to a determination result of whether the signal strength of the target area meets a preset high strength requirement, and determine target pose information of each of the multiple point clouds according to the point cloud matching pair, wherein there are identical points between two point clouds constituting the point cloud matching pair;

[0116] The generating module 504 is configured to generate a map of the target area according to the target pose information of each of the multiple frame point clouds.

[0117] Through the above technical solution, the multi-frame point cloud collected by the point cloud acquisition device in the target area is obtained, and it is determined whether the signal strength of the target area meets the preset high-intensity requirement. According to the determination result, the point cloud matching pair is determined from the multi-frame point cloud, and the target pose information of each of the multi-frame point clouds is determined according to the point cloud matching pair, and the map of the target area is generated according to the target pose information of each of the multi-frame point clouds. In this way, by constructing the point cloud matching pair and determining the target pose information of each of the multi-frame point clouds according to the point cloud matching pair, there is no need to directly obtain the pose information of the point cloud acquisition device according to the pose information of the global positioning system and the inertial measurement unit, which reduces the dependence on the global positioning system, and can generate a map with higher accuracy even when the signal of the global positioning system is weak. In addition, if there are identical points between two frames of point clouds, these two frames of point clouds can construct point cloud matching pairs. There is no need to find the nearest point to match each point as in the related art. The amount of calculation is relatively small. Based on the determination result of whether the signal strength of the target area meets the preset high-intensity requirement, point cloud matching pairs are determined from multiple frames of point clouds. This allows the number of constructed point cloud matching pairs to be adapted to the signal strength of the target area. By constructing point cloud matching pairs, the efficiency and accuracy of map generation can be improved.

[0118] Optionally, when the determination result indicates that the signal strength of the target area meets a preset high-intensity requirement, the second determination module 503 is used to determine a point cloud matching pair from the multi-frame point cloud in the following manner: for each frame point cloud, other frame point clouds in the multi-frame point cloud that have the same points as the frame point cloud and have the farthest timestamp distance from the frame point cloud constitute the point cloud matching pair with the frame point cloud.

[0119] Optionally, when the determination result indicates that the signal strength of the target area meets a preset high strength requirement, the second determination module 503 includes:

[0120] A first determination submodule is configured to determine, for each point cloud matching pair, respective spliced ​​point clouds of two frames of point clouds constituting the point cloud matching pair, and determine relative pose information between the two frames of point clouds based on the respective spliced ​​point clouds of the two frames of point clouds, wherein the spliced ​​point cloud of each frame of point cloud is spliced ​​by point clouds collected by the point cloud acquisition device in a specified area, and the specified area is an area around a position where the point cloud acquisition device collects the frame of point cloud;

[0121] The second determination submodule is configured to determine the target pose information of each of the multiple frames of point clouds according to the relative pose information corresponding to each point cloud matching pair.

[0122] Optionally, when the determination result indicates that the signal strength of the target area does not meet the preset high intensity requirement, the second determination module 503 is used to determine a point cloud matching pair from the multi-frame point cloud in the following manner: for each frame point cloud, other frame point clouds in the multi-frame point cloud that have the same points as the frame point cloud are combined with the frame point cloud to form the point cloud matching pair.

[0123] Optionally, when the determination result indicates that the signal strength of the target area does not meet a preset high strength requirement, the second determination module 503 includes:

[0124] The third determination submodule is configured to determine, for each of the point cloud matching pairs, the feature points of the two frames of point clouds constituting the point cloud matching pair, and determine the same-name points in the two frames of point clouds from the feature points; with the goal of minimizing the distance information between the same-name points, iteratively adjust the current posture information of the two frames of point clouds until the distance information between the same-name points is less than a preset distance threshold, thereby obtaining the target posture information of the two frames of point clouds.

[0125] Optionally, the first determining module 502 includes:

[0126] A fourth determination submodule is configured to determine first device posture information of each first preset time period during the process of the point cloud acquisition device acquiring the multiple frames of point clouds in the target area, and first confidences corresponding to each of the first device posture information;

[0127] The fifth determination submodule is configured to determine whether the signal strength of the target area meets a preset high strength requirement according to the first confidence level.

[0128] Optionally, the fourth determining submodule includes:

[0129] an acquisition submodule, configured to acquire second device posture information of the navigation device collected by the navigation device in the target area at second preset time intervals, and second confidences corresponding to each piece of the second device posture information;

[0130] a sixth determination submodule, configured to determine the first device posture information according to the relative position relationship between the navigation device and the point cloud acquisition device, and the second device posture information;

[0131] The seventh determination submodule is configured to determine, for each first device posture information, the target second device posture information in the second device posture information that is closest to the timestamp of the first device posture information, and use the second confidence of the target second device posture information as the first confidence corresponding to the first device posture information.

[0132] Optionally, the fifth determination module is used to: if the proportion of the first confidence levels higher than a preset confidence threshold is greater than a preset proportion threshold, or the number of first confidence levels higher than the preset confidence threshold is greater than a preset number threshold, then determine that the signal strength of the target area meets the preset high intensity requirement.

[0133] Regarding the device in the above embodiment, the specific manner in which each module performs operations has been described in detail in the embodiment of the method, and will not be elaborated here.

[0134] Figure 6 FIG. 7 is a block diagram of an electronic device 700 according to an exemplary embodiment. Figure 6 As shown, the electronic device 700 may include: a processor 701 and a memory 702. The electronic device 700 may also include one or more of a multimedia component 703, an input / output (I / O) interface 704, and a communication component 705.

[0135] The processor 701 is used to control the overall operation of the electronic device 700 to complete all or part of the steps in the above-mentioned map generation method. The memory 702 is used to store various types of data to support the operation of the electronic device 700. For example, these data may include instructions for any application or method used to operate on the electronic device 700, and application-related data, such as contact data, sent and received messages, pictures, audio, video, etc. The memory 702 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (Static Random Access Memory, referred to as SRAM), electrically erasable programmable read-only memory (Electrically Erasable Programmable Read-Only Memory, referred to as EEPROM), erasable programmable read-only memory (Erasable Programmable Read-Only Memory, referred to as EPROM), programmable read-only memory (Programmable Read-Only Memory, referred to as PROM), read-only memory (Read-Only Memory, referred to as ROM), magnetic memory, flash memory, magnetic disk or optical disk. The multimedia component 703 may include a screen and an audio component. The screen may be, for example, a touch screen, and the audio component is used to output and / or input audio signals. For example, the audio component may include a microphone for receiving external audio signals. The received audio signal may be further stored in the memory 702 or sent through the communication component 705. The audio component also includes at least one speaker for outputting audio signals. The I / O interface 704 provides an interface between the processor 701 and other interface modules, and the other interface modules may be keyboards, mice, buttons, etc. These buttons may be virtual buttons or physical buttons. The communication component 705 is used for wired or wireless communication between the electronic device 700 and other devices. Wireless communication, such as Wi-Fi, Bluetooth, Near Field Communication (NFC), 2G, 3G, 4G, NB-IOT, eMTC, or other 5G, etc., or a combination of one or more of them, is not limited here. Therefore, the corresponding communication component 705 may include: Wi-Fi module, Bluetooth module, NFC module, etc.

[0136] In an exemplary embodiment, the electronic device 700 can be implemented by one or more application specific integrated circuits (ASIC), digital signal processors (DSP), digital signal processing devices (DSPD), programmable logic devices (PLD), field programmable gate arrays (FPGA), controllers, microcontrollers, microprocessors or other electronic components to execute the above-mentioned map generation method.

[0137] In another exemplary embodiment, a computer-readable storage medium including program instructions is also provided, and when the program instructions are executed by a processor, the steps of the above-mentioned map generation method are implemented. For example, the computer-readable storage medium can be the above-mentioned memory 702 including program instructions, and the above-mentioned program instructions can be executed by the processor 701 of the electronic device 700 to complete the above-mentioned map generation method.

[0138] Figure 7 1 is a block diagram of an electronic device 1900 according to an exemplary embodiment. For example, the electronic device 1900 may be provided as a server. Figure 7 The electronic device 1900 includes a processor 1922, which may be one or more, and a memory 1932 for storing a computer program executable by the processor 1922. The computer program stored in the memory 1932 may include one or more modules each corresponding to a set of instructions. In addition, the processor 1922 may be configured to execute the computer program to perform the above-mentioned map generation method.

[0139] In addition, the electronic device 1900 may further include a power supply component 1926 and a communication component 1950. The power supply component 1926 may be configured to perform power management of the electronic device 1900, and the communication component 1950 may be configured to implement communication of the electronic device 1900, for example, wired or wireless communication. In addition, the electronic device 1900 may further include an input / output (I / O) interface 1958. The electronic device 1900 may operate based on an operating system stored in the memory 1932, such as Windows Server 2003. TM , Mac OS X TM , Unix TM , Linux TM etc.

[0140] In another exemplary embodiment, a computer-readable storage medium including program instructions is also provided, and when the program instructions are executed by a processor, the steps of the above-mentioned map generation method are implemented. For example, the computer-readable storage medium can be the above-mentioned memory 1932 including program instructions, and the above-mentioned program instructions can be executed by the processor 1922 of the electronic device 1900 to complete the above-mentioned map generation method.

[0141] In another exemplary embodiment, a computer program product is also provided. The computer program product includes a computer program executable by a programmable device, and the computer program has a code portion for executing the above-mentioned map generation method when executed by the programmable device.

[0142] The preferred embodiments of the present disclosure are described in detail above in conjunction with the accompanying drawings; however, the present disclosure is not limited to the specific details in the above embodiments. Within the technical concept of the present disclosure, a variety of simple modifications can be made to the technical solution of the present disclosure, and these simple modifications all fall within the protection scope of the present disclosure.

[0143] It should also be noted that the various specific technical features described in the above specific embodiments can be combined in any suitable manner without contradiction. In order to avoid unnecessary repetition, the present disclosure will not further describe various possible combinations.

[0144] In addition, various embodiments of the present disclosure may be arbitrarily combined, and as long as they do not violate the concept of the present disclosure, they should also be regarded as the contents disclosed by the present disclosure.

Claims

1. A map generation method, characterized in that: The method comprises: Acquire multiple frames of point clouds collected by a point cloud collection device in a target area; Determining whether the signal strength of the target area meets a preset high strength requirement; Determining a point cloud matching pair from the multiple point clouds according to a result of determining whether the signal strength of the target area meets a preset high-intensity requirement, and determining target pose information of each of the multiple point clouds according to the point cloud matching pair, wherein there are identical points between two point clouds constituting the point cloud matching pair; In the case where the determination result indicates that the signal strength of the target area does not meet the preset high strength requirement, determining the target pose information of each of the multiple frame point clouds according to the point cloud matching pair includes: For each point cloud matching pair, the feature points of the two frames of point clouds constituting the point cloud matching pair are determined, and the same-name points in the two frames of point clouds are determined from the feature points; with the goal of minimizing the distance information between the same-name points, the current pose information of the two frames of point clouds is iteratively adjusted until the distance information between the same-name points is less than a preset distance threshold, thereby obtaining the target pose information of the two frames of point clouds; A map of the target area is generated according to the target pose information of each of the multiple frame point clouds.

2. The method according to claim 1, characterized in that: In the case where the determination result indicates that the signal strength of the target area meets the preset high strength requirement, determining a point cloud matching pair from the multi-frame point cloud includes: For each frame point cloud, other frame point clouds in the multiple frame point clouds that have the same points as the frame point cloud and have the farthest timestamp distance from the frame point cloud form the point cloud matching pair with the frame point cloud.

3. The method according to claim 1, characterized in that In the case where the determination result indicates that the signal strength of the target area meets the preset high strength requirement, determining the target pose information of each of the multiple frame point clouds according to the point cloud matching pair includes: For each of the point cloud matching pairs, determine the spliced ​​point clouds of the two frames of point clouds constituting the point cloud matching pair, and determine the relative pose information between the two frames of point clouds based on the spliced ​​point clouds of the two frames of point clouds, wherein the spliced ​​point cloud of each frame of point cloud is formed by splicing point clouds collected by the point cloud acquisition device in a specified area, and the specified area is an area around the position where the point cloud acquisition device collects the frame of point cloud; The target pose information of each of the multiple frame point clouds is determined according to the relative pose information corresponding to each point cloud matching pair.

4. The method according to claim 1, characterized in that: In the case where the determination result indicates that the signal strength of the target area does not meet the preset high strength requirement, determining a point cloud matching pair from the multi-frame point cloud includes: For each frame point cloud, other frame point clouds in the multiple frame point clouds that have the same points as the frame point cloud are combined with the frame point cloud to form the point cloud matching pair.

5. The method according to claim 1, characterized in that The determining whether the signal strength of the target area meets the preset high strength requirement includes: Determine first device pose information at first preset time intervals during the process of the point cloud acquisition device acquiring the multiple frames of point clouds in the target area, and first confidences corresponding to each piece of the first device pose information; It is determined, according to the first confidence level, whether the signal strength of the target area meets a preset high strength requirement.

6. The method according to claim 5, characterized in that The determining of the first device posture information at intervals of a first preset time period during the process of the point cloud acquisition device acquiring the multiple frames of point clouds in the target area, and the first confidence corresponding to each of the first device posture information, includes: Acquire second device position information of the navigation device collected by the navigation device in the target area at intervals of a second preset time, and second confidences corresponding to each piece of the second device position information; Determine the first device posture information according to the relative position relationship between the navigation device and the point cloud acquisition device, and the second device posture information; For each first device posture information, determine the target second device posture information in the second device posture information that is closest to the timestamp of the first device posture information, and use the second confidence of the target second device posture information as the first confidence corresponding to the first device posture information.

7. The method according to claim 5, characterized in that The determining, according to the first confidence level, whether the signal strength of the target area meets a preset high strength requirement includes: If the proportion of the first confidences above the preset confidence threshold is greater than the preset proportion threshold, or the number of the first confidences above the preset confidence threshold is greater than the preset number threshold, it is determined that the signal strength of the target area meets the preset high strength requirement.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of the method described in any one of claims 1 to 7 are implemented.

9. An electronic device, characterized in that: include: a memory having a computer program stored thereon; A processor, configured to execute the computer program in the memory to implement the steps of the method according to any one of claims 1 to 7.

Images