Mechanical lidar control method and apparatus

Abstract

The application relates to a mechanical laser radar control method and device, computer equipment and a storage medium. The method comprises the following steps: determining at least one region of interest according to historical point cloud data; determining the number of detection modules corresponding to any region of interest; setting the detection region of each detection module according to the number of detection modules corresponding to each region of interest; and controlling each detection module to be turned on in the detection region corresponding to each detection module. The method can realize point cloud encryption of the region of interest in the detection range of the laser radar, and improve the energy utilization efficiency of the laser radar.

Description

Technical Field

[0001] This application relates to the field of lidar technology, and in particular to a mechanical lidar control method and device. Background Technology

[0002] With the development of autonomous driving technology, LiDAR has been rapidly popularized due to its high-precision and high-resolution perception capabilities; the usage rate of LiDAR in autonomous driving, assisted driving, and intelligent driving systems is rapidly increasing, and it is gradually entering the practical application stage.

[0003] Mechanical lidar is a type of radar that is widely used in specific scenarios such as robots, data testing in fixed scenarios, roadside vehicle-road cooperation, unmanned logistics vehicles, delivery vehicles, and Robo-taxi (autonomous taxis). It can provide a 360° field of view in the horizontal direction, enabling complete perception and modeling of the real physical world.

[0004] Traditional 360° horizontal field-of-view mechanical lidar can generally only achieve uniform or identical point cloud densification within the horizontal field of view. If it is necessary to improve the imaging resolution of certain regions of interest by densifying the point cloud, it is necessary to densify the point cloud of the entire field of view. However, densifying the point cloud means a greater power output. Using this method to improve the imaging resolution of the region of interest will result in more energy being wasted in non-regions of interest, leading to low energy utilization of the lidar. Summary of the Invention

[0005] Therefore, it is necessary to provide a mechanical lidar control method, device, computer equipment, and storage medium to address the aforementioned technical problems.

[0006] Firstly, this application provides a mechanical lidar control method. The method includes:

[0007] Based on historical point cloud data, identify at least one region of interest;

[0008] For any of the regions of interest, determine the number of detection modules corresponding to the region of interest;

[0009] Based on the number of detection modules corresponding to each region of interest, the detection area of ​​each detection module is set respectively. For any region of interest, the detection areas of each detection module corresponding to the region of interest overlap in the region of interest.

[0010] Each of the aforementioned detection modules is controlled to activate within its corresponding detection area.

[0011] In one embodiment, determining the number of detection modules corresponding to any given region of interest includes:

[0012] Acquire at least one historical radar detection frame and determine the historical region of interest corresponding to each historical radar detection frame;

[0013] The region of interest is matched with each of the historical regions of interest to obtain the target historical regions of interest that match the region of interest, and the number of historical detection modules corresponding to the target historical regions of interest is determined.

[0014] Based on the historical number of detection modules and the preset strategy, the number of detection modules for the region of interest is set.

[0015] In one embodiment, setting the detection area of ​​each detection module according to the number of detection modules corresponding to each region of interest includes:

[0016] Determine the target segmentation region corresponding to each of the regions of interest in each segmentation region. The segmentation region is obtained by dividing the detection range of the lidar, and there is no overlap between the segmentation regions.

[0017] Based on the relative positions between the target segmentation regions, the detection area and phase of each detection module are set respectively.

[0018] In one embodiment, when both the number of detection modules and the number of regions of interest are two, the segmented regions include four, and each segmented region has a field of view of 90°; the step of setting the detection region and phase of each detection module according to the relative position between each target segmented region includes:

[0019] When the two regions of interest are respectively included in the two target segmentation regions and the field of view of each region of interest does not exceed 90°, the detection area size of the two detection modules is set between 180° and 360°, and the phase of the two detection modules is set so that the phase difference between the two detection modules is between 90° and 180°.

[0020] In one embodiment, when the two target segmentation regions are symmetrical about the radar center, the two boundaries of one region of interest are the start position of the detection region of one detection module and the end position of the detection region of the other detection module, respectively.

[0021] The two boundaries of the other region of interest are the start position of the detection region of one detection module and the end position of the detection region of the other detection module, respectively.

[0022] The detection area of ​​both detection modules is greater than 180°.

[0023] In one embodiment, when the number of detection modules and regions of interest are both 2, the segmented regions include 4, and the field of view of each segmented region is 90°; the step of setting the detection region and phase of each detection module according to the relative position between each target segmented region includes:

[0024] When the field of view of at least one of the regions of interest is greater than 90°, the size of the detection area of ​​at least one of the detection modules is set between 270° and 360°.

[0025] Secondly, this application also provides a mechanical lidar control device. The device includes:

[0026] The first determination module is used to determine at least one region of interest based on historical point cloud data;

[0027] The second determining module is used to determine the number of detection modules corresponding to any given region of interest.

[0028] The setting module is used to set the detection area of ​​each detection module according to the number of detection modules corresponding to each region of interest. For any region of interest, the detection areas of each detection module corresponding to the region of interest overlap in the region of interest.

[0029] A control module is used to control each of the detection modules to be activated within the detection area corresponding to each of the detection modules.

[0030] In one embodiment, the second determining module is further configured to:

[0031] Acquire at least one historical radar detection frame and determine the historical region of interest corresponding to each historical radar detection frame;

[0032] The region of interest is matched with each of the historical regions of interest to obtain the target historical regions of interest that match the region of interest, and the number of historical detection modules corresponding to the target historical regions of interest is determined.

[0033] Based on the historical number of detection modules and the preset strategy, the number of detection modules for the region of interest is set.

[0034] In one embodiment, the setting module is further configured to:

[0035] Determine the target segmentation region corresponding to each of the regions of interest in each segmentation region. The segmentation region is obtained by dividing the detection range of the lidar, and there is no overlap between the segmentation regions.

[0036] Based on the relative positions between the target segmentation regions, the detection area and phase of each detection module are set respectively.

[0037] In one embodiment, when the number of detection modules and regions of interest are both 2, the segmented regions include 4, and the field of view of each segmented region is 90°;

[0038] The setting module is further configured to, when the two regions of interest are respectively included in the two target segmentation regions and the field of view of each region of interest does not exceed 90°, set the detection area size of the two detection modules to be between 180° and 360°, and set the phase of the two detection modules so that the phase difference between the two detection modules is between 90° and 180°.

[0039] In one embodiment, when the two target segmentation regions are symmetrical about the radar center, the two boundaries of one region of interest are the start position of the detection region of one detection module and the end position of the detection region of the other detection module, respectively.

[0040] The two boundaries of the other region of interest are the start position of the detection region of one detection module and the end position of the detection region of the other detection module, respectively.

[0041] The detection area of ​​both detection modules is greater than 180°.

[0042] In one embodiment, when the number of detection modules and regions of interest are both 2, the segmented regions include 4, and the field of view of each segmented region is 90°; the step of setting the detection region and phase of each detection module according to the relative position between each target segmented region includes:

[0043] When the field of view of at least one of the regions of interest is greater than 90°, the size of the detection area of ​​at least one of the detection modules is set between 270° and 360°.

[0044] Thirdly, this application also provides a computer device. The computer device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement any of the methods described above.

[0045] Fourthly, this application also provides a computer-readable storage medium. The computer-readable storage medium stores a computer program thereon, which, when executed by a processor, implements any of the above methods.

[0046] Fifthly, this application also provides a computer program product. The computer program product includes a computer program that, when executed by a processor, implements any of the above methods.

[0047] The aforementioned mechanical lidar control method, device, computer equipment, and storage medium determine the region of interest (ROI) based on historical point cloud data. Then, based on the number of detection modules within the ROI, the detection area of ​​each detection module is set, enabling point cloud encryption within the ROI. This embodiment, by setting multiple detection modules, allows for real-time dynamic adjustment of the ROI during lidar operation, based on the determined ROI, thereby improving the lidar's energy utilization rate. Attached Figure Description

[0048] Figure 1 This is a diagram illustrating the application environment of a mechanical lidar control method in one embodiment.

[0049] Figure 2 This is a flowchart illustrating a mechanical lidar control method in one embodiment;

[0050] Figure 3 This is a schematic diagram of the region of interest in one embodiment;

[0051] Figure 4 This is a schematic diagram illustrating the detection timing of the detection module in one embodiment;

[0052] Figure 5 This is a flowchart illustrating step 202 in one embodiment;

[0053] Figure 6 This is a schematic diagram illustrating the construction of the region of interest in one embodiment;

[0054] Figure 7 This is a flowchart illustrating step 202 in one embodiment;

[0055] Figure 8 This is a flowchart illustrating step 204 in one embodiment;

[0056] Figure 9 This is a flowchart illustrating step 206 in one embodiment;

[0057] Figure 10 This is a schematic diagram of a segmented region in one embodiment;

[0058] Figure 11This is a schematic diagram showing the relative positions of different target segmentation regions and their corresponding detection regions in one embodiment;

[0059] Figure 12 This is a schematic diagram illustrating the setting of a detection region based on the angle between the boundary of the region of interest and the termination line of the target segmentation region in one embodiment.

[0060] Figure 13 This is a schematic diagram illustrating the setting of a detection region according to the boundary of the region of interest in one embodiment.

[0061] Figure 14 This is a schematic diagram of setting up a detection area in one embodiment when the field of view of the region of interest is greater than 90°;

[0062] Figure 15 This is a structural block diagram of a mechanical lidar control device in one embodiment;

[0063] Figure 16 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0064] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0065] In one embodiment, such as Figure 1 As shown, the mechanical lidar control method provided in this application embodiment can be applied to, but is not limited to, [examples of other methods]. Figure 1In the application environment shown, the control module 102 communicates with the laser emitting module 104 and at least two detection modules 106. Before the lidar begins detection in the current frame, the control module 102 determines the region of interest (ROI) based on historical point cloud data and the number of detection modules for each ROI. Then, by setting the detection area of ​​each detection module 106, each ROI corresponds to at least one detection module 106, and the detection areas of each detection module 106 overlap within the ROI. During the lidar's detection in the current frame, the control module 102 controls the laser emitting module 104 to emit laser light, and simultaneously controls each detection module 106 to activate within its respective detection area, so that the detection module 106 can receive laser echoes within its detection area. The laser emitting module 104 may be, but is not limited to, a Vcsel (Vertical Cavity Surface Emitting Laser), an EEL (Edge-Emitting Laser), or an LD (Laser Diode), etc. The detection module 106 may be, but is not limited to, a single-photon detector array (SPAD), a conventional APD (Avalanche Photodiode), PD (Photodiode), SIPM (Silicon Photomultiplier), etc., and their single-point mounting, array, or other forms.

[0066] In one embodiment, such as Figure 2 As shown, a mechanical lidar control method is provided. This embodiment applies this method to... Figure 1 Taking the control module 102 as an example, the explanation includes the following steps:

[0067] Step 202: Based on historical point cloud data, determine at least one region of interest.

[0068] In this embodiment, historical point cloud data refers to point cloud data acquired by the LiDAR during historical detection. Based on this historical point cloud data, environmental information surrounding the LiDAR can be determined, thereby identifying the region of interest (ROI). Those skilled in the art can pre-set ROI determination criteria and incorporate these criteria into the control module, enabling the control module to automatically determine the ROI during LiDAR operation. For example, when LiDAR is applied to autonomous driving, those skilled in the art can define the road directly ahead of the vehicle as the ROI. The control module can determine which point data represents the road ahead based on the historical point cloud data, and then determine the ROI based on the relative positions of these point data and the LiDAR.

[0069] Step 204: For any region of interest, determine the number of detection modules corresponding to the region of interest.

[0070] In this embodiment, multiple detection modules can be set up to receive laser echoes in the region of interest, thereby increasing the number of point data detected by the lidar in the region of interest. This is equivalent to encrypting the data in the region of interest and improving the imaging resolution of the region of interest.

[0071] Since the number of detection modules cannot exceed the total number of detection modules in the lidar, the control module should first obtain the total number of detection modules in the lidar, then determine the selectable range of detection modules corresponding to the region of interest based on the total number, and then determine the number of detection modules for each region of interest from the selectable range.

[0072] Step 206: Based on the number of detection modules corresponding to each region of interest, set the detection area of ​​each detection module. For any region of interest, the detection areas of each detection module corresponding to the region of interest overlap in the region of interest.

[0073] In this embodiment of the application, after determining the number of detection modules corresponding to each region of interest, the detection area corresponding to each detection module in the lidar can be set according to the number of detection modules, so that for each region of interest, there is a detection area of ​​the number of detection modules corresponding to that region of interest, which overlaps in that region of interest.

[0074] For example, such as Figure 3 As shown, the gray area represents the region of interest, which is symmetrical about the center of the radar circle. Each region of interest corresponds to two detection modules. When the scanning module scans the gray area, the two detection modules corresponding to that area are activated simultaneously to increase the number of photons received and improve the sensing effect.

[0075] Step 208: Control each detection module to activate it within its corresponding detection area.

[0076] In this embodiment of the application, after the control module determines the detection area corresponding to each detection module, the control module can determine the detection timing of each detection module based on the detection start position of each frame of the lidar and the area where each detection area is located relative to the detection start position, that is, the on-time and off-time in a frame of detection, so that each detection module can be turned on in the set detection area and turned off in other areas.

[0077] Since the scanning frequency of a lidar is constant, its angular velocity of rotation is generally uniform and does not change during operation. Therefore, the specific time it takes for the lidar to rotate to each angle can be determined. For example, if the lidar's scanning frequency is 10Hz, meaning it takes 0.1 seconds to complete one scan, then it can be determined that when the lidar rotates to a 90° azimuth position, it is 0.025 seconds after the lidar begins its current frame of detection; when it rotates to a 180° azimuth position, it is 0.05 seconds after the lidar begins its current frame of detection, and the corresponding times for other angles can be deduced similarly.

[0078] The angle between the starting boundary and the initial detection position of the detection area, and the angle between the ending boundary and the initial detection position of the detection area, can be determined. Based on these two angles and the scanning frequency of the lidar, the detection timing of the detection module can be determined. (Refer to...) Figure 4 As shown, if the dashed line represents the detection start position and the shaded area represents the detection area, the angle between the start boundary of the detection area and the detection start position is 0°, the angle between the end boundary of the detection area and the detection start position is 225°, and the scanning frequency of the lidar is 10Hz, then it can be determined that when the lidar rotates to the 0° position, the corresponding time is 0 seconds after the start of the current frame scan, and when the lidar rotates to the 225° position, the corresponding time is 0.0625 seconds after the start of the current frame scan. Therefore, it can be set that the detection module should be turned on at 0 seconds after the start of the current frame scan and turned off after 0.0625 seconds, so that the detection module can be turned on within its corresponding detection area.

[0079] The mechanical lidar control method provided in this application determines the region of interest (ROI) based on historical point cloud data. Then, based on the number of detection modules within the ROI, the detection area of ​​each module is set, thereby achieving point cloud encryption within the ROI and resulting in better sensing performance. By setting multiple detection modules, this application allows for real-time dynamic adjustment of the ROI during lidar operation, based on the determined ROI, thus improving the lidar's energy utilization rate.

[0080] In one embodiment, such as Figure 5 As shown, in step 202, at least one region of interest is determined based on historical point cloud data, including:

[0081] Step 502: Perform object recognition on the historical point cloud data and obtain the recognition results.

[0082] Step 504: If a target recognition result exists in the recognition results, construct at least one region of interest based on the relative position of the target recognition result and the lidar.

[0083] In this embodiment, based on historical point cloud data, various objects within the LiDAR detection area can be segmented and identified, i.e., the identification results. The target identification results are one or more of the identification results that the LiDAR needs to focus on detecting. This embodiment does not specifically limit the method of obtaining identification results from historical point cloud data; any point cloud object recognition algorithm for LiDAR is applicable to this embodiment.

[0084] Depending on the application environment, the objects that a LiDAR needs to focus on detecting differ, and therefore the target recognition results may also vary accordingly. Target recognition results can be manually set by someone skilled in the art, or the control module can adaptively adjust them based on the current scene. If manual setting by someone skilled in the art is required, the target recognition results can be written into the control module's memory before the LiDAR starts operating; alternatively, during LiDAR operation, the target recognition results written into the memory can be added to or deleted via an external control device to adjust the target recognition results. If the control module needs to adaptively adjust the target recognition results based on the LiDAR's usage scenario, the correspondence between the scene and the target recognition results can be written into the control module's memory, allowing the control module to automatically detect the current scene and determine the target recognition results based on the correspondence between the scene and the memory.

[0085] For example, when LiDAR is applied to the field of autonomous driving, two correspondences between scenes and target recognition results can be established: when the car is driving on the road, the target recognition result is the road ahead and other vehicles relatively close to the car; when the car is passing through an intersection, the target recognition result is other vehicles, non-motorized vehicles, and pedestrians ahead. The LiDAR control module can obtain the current scene from the car's autonomous driving control module (the car's autonomous driving module can determine whether the car is currently on the road or at an intersection through the onboard map and the car's current location, thereby determining the current scene); or, the control module can also determine the current scene based on the point cloud data collected by the LiDAR. For example, if object recognition is performed on the point cloud data, and the point cloud data contains point data representing traffic lights, then the current scene can be determined to be the car passing through an intersection. This application embodiment does not specifically limit this.

[0086] After determining the current scene, the control module searches for corresponding target recognition results based on various relationships. For example, if the current scene is a car driving on a road, the control module can find the target recognition results corresponding to this scene as the road ahead and other vehicles relatively close to the vehicle. The control module can then use the road ahead and other vehicles relatively close to the vehicle as target recognition results.

[0087] After determining the target recognition result and detecting the presence of a target recognition result among the various recognition results, a region of interest (ROI) can be constructed based on the relative position of the target recognition result and the LiDAR. If neither the target recognition result nor the LiDAR moves, the ROI can be directly constructed based on the position of the target recognition result relative to the LiDAR in historical point cloud data.

[0088] Since the target recognition result may be a moving object, and the LiDAR is mounted on a movable device (such as a car, robot, etc.), the LiDAR may also move, thus changing the relative position between the target recognition result and the LiDAR. If at least one of the target recognition result or the LiDAR moves, the control module needs to predict the relative position between them in the next frame scan based on the moving speed of the target recognition result and the moving speed of the LiDAR, thereby constructing the region of interest in the next frame scan. The moving speed of the target recognition result can be determined based on historical point cloud data obtained from past multi-frame scans, while the moving speed of the LiDAR (i.e., the moving speed of the device on which the LiDAR is located) can be obtained through communication between the LiDAR and its device; this embodiment does not specifically limit this.

[0089] by Figure 6 To illustrate with a simple example, if the target recognition result and the LiDAR move towards each other at a relative speed of 30 m / s, and the LiDAR completes one frame scan in 0.1 seconds, and in the previous frame of historical point cloud data, the target recognition result is located at the 315° azimuth of the LiDAR, and the distance between the target recognition result and the LiDAR is 4.2 m, then it can be predicted that in the next frame scan, the target recognition result will be approximately located at the 270° azimuth of the LiDAR. Therefore, a region of interest (ROI) can be constructed based on the predicted location of the target recognition result (270° azimuth). The method of constructing the ROI based on the predicted location of the target recognition result can refer to the method of constructing the ROI based on the actual location of the target recognition result described above; this embodiment will not be repeated here.

[0090] The mechanical lidar control method provided in this application determines whether target recognition results exist within the lidar's detection range based on historical point cloud data. If target recognition results exist, a region of interest is constructed based on the relative position of the target recognition results and the lidar. Therefore, when target recognition results requiring focused detection exist, point cloud encryption can be applied to the area where the target recognition results are located, thereby improving the imaging effect in that area. This allows the lidar to increase energy output specifically for that area, improving the lidar's energy utilization rate.

[0091] In one embodiment, such as Figure 7As shown, in step 202, at least one region of interest is determined based on historical point cloud data, including:

[0092] Step 702: Obtain multiple historical radar detection frames and the historical point cloud data corresponding to each historical radar detection frame.

[0093] Step 704: For any historical radar detection frame, detect sparse points in the historical point cloud data corresponding to the historical radar detection frame.

[0094] Step 706: For any target sparse point, match each historical radar detection frame with the target sparse point according to the sparse point corresponding to each historical radar detection frame. If a preset number of historical radar detection frames match the target sparse point, construct at least one region of interest according to the relative position of the target sparse point and the lidar. The target sparse point is any sparse point.

[0095] In this embodiment, sparse points refer to a set of point data that are spaced far apart. When historical point cloud data shows that sparse points continuously appear at the same location, it indicates that there may be an object at that location. However, because the point data collected at this location is relatively sparse, the detection accuracy of the lidar for this object is not high. In this case, the area where the sparse points are located can be designated as the region of interest to increase the number of point data collected in that area.

[0096] Before a lidar detection frame begins, multiple historical lidar detection frames preceding the current frame can be acquired (the specific number can be determined by those skilled in the art according to actual needs, such as 5 or 10), and historical point cloud data collected in these historical lidar detection frames can be acquired. For any historical lidar detection frame, sparse points in the historical lidar detection frame can then be detected. For example, point data that represents the same object but is sparsely distributed can be considered as sparse points: that is, the historical point cloud data is clustered, and point data belonging to the same point cloud cluster but whose distance to each other exceeds a distance threshold (the distance threshold can be determined by those skilled in the art) are considered as sparse points; or, point cloud clusters that cannot be identified in the historical lidar detection frame can also be considered as sparse points: that is, after each frame of point cloud data is acquired and object recognition is performed on the point cloud data, point data belonging to the same point cloud cluster but for which the object recognition result for the point cloud cluster cannot be determined are considered as sparse points. This application embodiment does not specifically limit this.

[0097] After identifying sparse points in each historical radar detection frame, these sparse points can be matched to determine which sparse points appear simultaneously in multiple historical radar detection frames. For any given sparse point, it can be designated as a target sparse point. Based on parameters such as the cluster shape of the point cloud cluster to which the target sparse point belongs, the moving speed of the lidar, the lidar's scanning frequency, and the moving speed of the target sparse point, matching is performed between sparse points in other historical radar detection frames and the target sparse point. When a preset number of historical radar detection frames and target sparse points are matched (the preset number can be determined by those skilled in the art based on actual needs, and its value should be less than the number of selected historical radar detection frames), it can be determined that an object exists at the location corresponding to the target sparse point. Therefore, a region of interest (ROI) can be constructed based on the relative position of the target sparse point and the lidar, allowing for focused detection at that location. The method of constructing the ROI based on the relative position of the target sparse point and the lidar can be found in the aforementioned embodiments, where the method of constructing the ROI based on the target recognition result and the relative position of the lidar is described again in this application.

[0098] The above process is illustrated with a specific example. After determining the sparse target points from any historical radar detection frame, the cluster shape of the sparse target points can be used to determine whether there are point cloud clusters (hereinafter referred to as matching point cloud clusters) in other historical radar detection frames with similar shapes. That is, the cluster shapes of the sparse target points and the point cloud clusters in each historical radar detection frame are matched. This application does not specifically limit the matching algorithm; any algorithm that can perform shape matching between two clustering results is applicable to this application.

[0099] The mechanical lidar control method provided in this application determines whether sparse points exist within the lidar's detection range based on historical point cloud data. When sparse points appear simultaneously in multiple historical lidar detection frames, a region of interest is constructed based on the relative positions of the sparse points and the lidar. Therefore, when there are sparse points requiring enhanced detection, the area where the sparse points are located can be densified, thereby improving the imaging effect in that area. Consequently, the lidar can increase its energy output specifically for that area, improving the lidar's energy utilization rate.

[0100] In one embodiment, such as Figure 8 As shown, step 204 includes:

[0101] Step 802: Acquire at least one historical radar detection frame and determine the historical region of interest corresponding to each historical radar detection frame.

[0102] Step 804: Match the region of interest with each historical region of interest to obtain the target historical region of interest that matches the region of interest, and determine the number of historical detection modules corresponding to the target historical region of interest.

[0103] Step 806: Based on the historical number of detection modules and the preset strategy, set the number of detection modules for the region of interest.

[0104] In this embodiment, after determining the region of interest (ROI) and the number of detection modules for that ROI each time, the ROI and the historical number of detection modules can be stored as configuration parameters for historical ROIs. This allows the number of detection modules required for the ROI to be determined after acquiring the ROI corresponding to the current detection frame, based on the target historical ROIs that match the ROI in historical radar detection frames and the number of historical detection modules used for the target historical ROIs. If there are not enough historical radar detection frames currently available, the detection module can use a preset number of detection modules as the number of detection modules corresponding to the ROI.

[0105] The target (including target recognition results and target sparse points) of each historical region of interest can be determined, and historical regions of interest that target the same target as the region of interest are designated as target historical regions of interest. There can be one or more target historical regions of interest. "Same target" can mean the same type of target (e.g., if the target recognition result is a vehicle in front, then all historical regions of interest constructed for the vehicle in front can be target historical regions of interest), or it can mean the same target (e.g., if the target recognition result is a specific vehicle, then historical regions of interest constructed for that vehicle are target historical regions of interest; historical regions of interest constructed for the same target sparse points are target historical regions of interest, etc.).

[0106] Alternatively, the parameters of the target targeted by each historical region of interest (ROI) can be determined, and the historical ROIs with similar parameters to those of the target targeted by the ROI can be used as the target's historical ROI. These parameters can include the distance between the target and the lidar, the angle relative to the lidar, etc. For example, if the target targeted by the ROI is located 10m from the lidar, then a historical ROI constructed based on a target approximately 10m from the lidar can be determined from the historical ROIs, and this historical ROI can then be used as the target's historical ROI.

[0107] In addition, the target historical region of interest can also be determined based on multiple criteria mentioned above. For example, a historical region of interest that is the same as the target targeted by the region of interest or has similar parameters to the target targeted by the region of interest can be used as the target historical region of interest. This application does not specifically limit this.

[0108] A preset strategy can refer to setting the number of detection modules in the region of interest (ROI) to be equal to the number of one historical detection module, or the number of detection modules to be greater than the total number of historical detection modules. This can be preset by those skilled in the art based on actual needs. For example, the preset strategy can be set so that the number of detection modules in the ROI is equal to the largest number of historical detection modules, ensuring that the LiDAR's imaging quality in the ROI is at least equal to the best imaging quality achieved during historical detection. Alternatively, the preset strategy can be set so that the number of detection modules in the ROI is less than the total number of historical detection modules, until the number of detection modules reaches a threshold, allowing the LiDAR's imaging quality to gradually improve.

[0109] Furthermore, the control module can adaptively adjust the number of detection modules based on whether the point cloud data acquired from historical detection module counts meets the requirements. To enable the control module to determine which target's historical region of interest (ROI) has a sufficient number of historical detection modules and how to set the laser detection module count based on historical counts, communication can be established between the control module and the processing device that processes the point cloud data acquired by the lidar. This allows the control module to receive feedback from the processing device on the acquired point cloud data. For example, after acquiring and sending point cloud data to the processing device in each frame, the processing device can send an indication signal to the control module based on whether the point cloud data density meets the requirements. The control module then stores the correspondence between the ROI, the number of detection modules, and the indication signal in the current frame.

[0110] When determining the number of probe modules for the region of interest in the next frame, the control module reads the number of historical probe modules and their indication signals. If the indication signals corresponding to all historical probe module counts indicate that the point cloud data density does not meet the requirements, the control module can set the number of probe modules to a value less than the total number of historical probe modules (according to a preset strategy, this can be set to one more or more historical probe modules than the minimum historical probe module count). If the indication signals corresponding to at least one historical probe module count indicate that the point cloud data density meets the requirements, the control module can set the number of probe modules to be equal to the number of one of the historical probe modules (according to a preset strategy, this can be set to be equal to the maximum historical probe module count or any historical probe module count). In this way, the control module can adaptively adjust the number of probe modules based on whether the historical probe module counts meet the requirements.

[0111] The mechanical lidar control method provided in this application determines the target's historical region of interest (ROI) that matches the region of interest (ROI), and sets the number of detection modules in the ROI based on the historical number of detection modules and preset strategies. Therefore, the number of detection modules required for this detection can be adaptively set according to the historical number of detection modules determined by the lidar in previous detections, thereby improving the accuracy of the detection module number setting.

[0112] In one embodiment, such as Figure 9 As shown, in step 206, the detection area of ​​each detection module is set according to the number of detection modules corresponding to each region of interest, including:

[0113] Step 902: Determine the target segmentation region corresponding to each region of interest in each segmentation region. The segmentation region is obtained by dividing the detection range of the lidar, and there is no overlap between the segmentation regions.

[0114] In this embodiment, the detection range of the lidar can be pre-divided into multiple non-overlapping segments. Each segment is a sector centered on the lidar's detection center (i.e., the center of the lidar's 360° detection range). Each segment can be a group of regions with the same central angle or a group of regions with different central angles. There are at least two regions of interest (ROIs). Based on the relative position of the ROIs to the lidar, the segmented region containing the ROIs is determined as the target segmented region.

[0115] Reference Figure 10 As shown, the radar's detection range is divided into four segments numbered 1 to 4. Each segment has a field of view of 90° and they do not overlap. The region of interest is the gray area, which is located in region 4. Therefore, the target segment for the region of interest is region 4.

[0116] Step 904: Based on the relative positions between the segmented regions of each target, set the detection area and phase of each detection module.

[0117] To ensure good detection results, low energy consumption, and relatively simple control logic, it is necessary to set up the detection areas of multiple detection modules and the phase differences between them. This ensures that the multiple detection modules are continuously turned on / off during the radar scanning process (uninterrupted operation) and cooperate with each other to complete the radar's 360° circumferential scan. They are also simultaneously turned on in the region of interest to perform point cloud encryption.

[0118] In one embodiment, when the total number of detection modules is 2 and the number of detection modules corresponding to the region of interest is also 2, the segmented region includes 4 regions, and the field of view of each segmented region is 90°. When the field of view of each region of interest does not exceed 90°, the detection area size of the two detection modules is between 180° and 270° and the phase difference is between 90° and 180°.

[0119] Since the field of view of each region of interest does not exceed 90°, each region can be included within a single segmented region. Based on the relative positions of the target segmented regions containing two regions of interest, they can be mainly categorized as adjacent or centrally symmetric. Figure 11 In (a), the target segmentation regions containing regions of interest S1 and S2 are symmetrical about the radar circle center. Figure 11 In (b), the target segmentation regions where regions of interest S3 and S4 are located are adjacent, and will be discussed separately below.

[0120] 1) The target segmentation region is centrally symmetric:

[0121] like Figure 11 (a) The field of view of regions of interest S1 and S2 does not exceed 90°, and the target segmentation regions where they are located are symmetrical about the center of the radar circle. In order to achieve densification of the target segmentation regions where the regions of interest are located, the detection area of ​​detection module 1 covers regions 1, 2, and 3; the detection area of ​​detection module 2 covers regions 3, 4, and 1. The detection area of ​​each detection module is 270°, and the phase difference between the two detection modules is 180°. At this time, the point cloud densification of target segmentation regions 1 and 3 can be achieved, and the working state of detection modules 1 and 2 is relatively continuous, without frequent opening or closing in a short period of time, reducing the complexity of control and energy consumption.

[0122] 2) Adjacent (asymmetric) target segmentation regions:

[0123] like Figure 11 (b) The field of view of regions of interest S3 and S4 does not exceed 90°, and their target segmentation regions are adjacent. In order to achieve densification of the target segmentation regions where the regions of interest are located, the detection area of ​​detection module 1 covers regions 1, 2, and 3; the detection area of ​​detection module 2 covers regions 2, 3, and 4. The detection area of ​​each detection module is 270°, and the phase difference between the two detection modules is 90°. At this time, the point cloud densification of target segmentation regions 2 and 3 can be achieved. The working states of detection modules 1 and 2 are relatively continuous, and there is no frequent opening or closing state in a short period of time, which reduces the complexity of control and energy consumption.

[0124] Thus, by employing two detection modules and considering the different relative positions of the two regions of interest within the target segmentation region, different cooperation mechanisms can be used to achieve symmetric or asymmetric encryption of the point cloud within the target segmentation region, as summarized in Table 1:

[0125] Table 1 Symmetric or Asymmetric Encryption

[0126]

[0127] As shown in Table 1, assuming the detection area of ​​detection module 1 is regions 1, 2, and 3, symmetrical or asymmetrical detection effects can be achieved by varying the ratio of detection areas in detection module 2. In Table 1, the detection areas of detection modules 1 and 2 are both 270°, but their phase differences are 90° or 180°.

[0128] In the above embodiments, in order to achieve 360° surround detection by the two detection modules, the minimum detection area of ​​each detection module should be 180°. In a further embodiment, to reduce energy consumption, the angle between the region of interest and the dividing line of each region can be calculated. On the basis of reducing the detection area of ​​each detection module to between 180° and 270°, the region of interest can be encrypted.

[0129] like Figure 12 (a) Assuming the radar scans in the order of region 1-2-3-4, the boundary lines of the four segmented regions are defined as the termination / starting lines of region X according to the radar's scanning direction. For example, the boundary line between region 1 and 4 is both the termination line of region 4 and the starting line of region 1.

[0130] 1) When the regions of interest S1 and S2 are located within two target segmentation regions that are symmetrical about the radar center, such as Figure 12 As shown in (a), taking S2 as an example, two angles are formed between the two boundaries of S2 and the termination line of region 1 (that is, the starting line of region 2). Calculate the smaller angle between the two angles formed by the boundary of region S2 and the termination line of region 1. Figure 12 In (a), the included angle is 2. Similarly, find the smaller angle between the two included angles formed by the boundary of another region of interest S1 and the termination line of its region 3, i.e. Figure 12 Angle 1 in (a).

[0131] It can be seen that although the regions corresponding to angles 2 and 1 are within the target segmentation region, they are not actually the locations of the region of interest. Therefore, at angles 2 and 1, two detection modules can be activated simultaneously, that is, if... Figure 12As shown in (a), for detection module 1, its detection area (dashed arc) can be set to (270° - angle 1), and similarly, for detection module 2, its detection area (solid arc) can be set to (270° - angle 2). The phase difference between the two detection modules remains 180°. Thus, while encrypting each region of interest, the power consumption of each detection module is further reduced.

[0132] Furthermore, due to Figure 12 In (a), there are still instances of energy waste, such as Figure 13 (a) and (b) present a more economical measure that eliminates the energy consumption corresponding to the non-interesting encrypted area. In this case, the starting working positions of the two probe modules no longer coincide with the dividing line between two regions. Figure 13 (a) One of the boundaries of regions of interest S1 and S2 is the starting point of the detection area of ​​the two detection modules, and the other boundary of regions of interest S1 and S2 is the ending point of the detection area of ​​the two detection modules.

[0133] Similarly, regarding Figure 13 (b) One of the boundaries of regions of interest S34 and S2 is the starting point of the detection area of ​​the two detection modules, and the other boundary of regions of interest S34 and S2 is the ending point of the detection area of ​​the two detection modules.

[0134] in, Figure 13 In (b), the region of interest S34 is greater than 90°, spanning regions 3 and 4, but still follows the above pattern. At this time, the detection regions of both detection modules are greater than 180°, but the phase difference is affected by the included angles 1 and 2, and is not a fixed value.

[0135] 2) When regions of interest S3 and S4 are located within two adjacent target segmentation regions, such as Figure 12 As shown in (b), the two adjacent target segmentation regions 2 and 3 are first stitched together to obtain a stitched region, which is a semicircle. According to the radar scanning direction, the start line and end line of the stitched region are determined, wherein the start line of the stitched region is the start line of region 2, and the end line of the stitched region is the end line of region 3.

[0136] Then, determine the smaller of the two included angles formed by the two boundaries of the region of interest S4 and the starting line of the splicing region, such as... Figure 12 (b) The smaller of the included angle 4, and the two included angles formed by the two boundaries of the region of interest S3 and the termination line of the splicing region, such as Figure 12 (b) includes the angle 3.

[0137] At this point, for detection module 1, its detection area is (270° - angle 3), and similarly, for detection module 2, its detection area is (270° - angle 4). The phase difference between the two detection modules is (90° + angle 4). Since angle 4 is always less than 90°, the phase difference between the two detection modules does not exceed 180°. Therefore, while encrypting each region of interest, the power consumption of each detection module is further reduced.

[0138] In other words, the mechanical lidar control method provided in this application, when there are 2 detection modules, 4 segmented regions, a field of view of 90° for each segmented region, and a field of view of each region of interest not exceeding 90°, sets the detection area size of the two detection modules between 180° and 270°, and the phase difference between 90° and 180°, so that the detection area of ​​the two detection modules is as small as possible in this case, saving the resources consumed by the lidar when scanning.

[0139] Of course, such as Figure 14 As shown, when the field of view of at least one region of interest S34 is greater than 90°, it indicates that the region of interest spans two segmented regions, and the field of view of at least one detection module (detection module 1) needs to be set between 270° and 360°. Figure 14 In (a), the detection area of ​​detection module 1 is greater than 270°. Figure 14 In (b), the detection areas of both detection modules are greater than 270°.

[0140] The mechanical lidar control method provided in this application, when there are 2 detection modules, 4 segmented regions, and a field of view of 90° for each segmented region, and a field of view of a region of interest exceeding 90°, sets the detection area size of at least one detection module between 270° and 360°, so that the detection area of ​​the two detection modules is as small as possible in this case, saving the resources consumed by the lidar when scanning.

[0141] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0142] Based on the same inventive concept, this application also provides a mechanical lidar control device for implementing the aforementioned mechanical lidar control method. The solution provided by this device is similar to the solution described in the above method; therefore, the specific limitations in one or more embodiments of the mechanical lidar control device provided below can be found in the limitations of the mechanical lidar control method described above, and will not be repeated here.

[0143] In one embodiment, such as Figure 15 As shown, a mechanical lidar control device is provided, comprising: a first determining module 1502, a second determining module 1504, a setting module 1506, and a control module 1508, wherein:

[0144] The first determining module 1502 is used to determine at least one region of interest based on historical point cloud data;

[0145] The second determining module 1504 is used to determine the region of interest for any given region of interest, and to determine the number of detection modules corresponding to the region of interest based on the given information.

[0146] Setting module 1506 is used to set the detection area of ​​each detection module according to the number of detection modules corresponding to each region of interest, and for any region of interest, the detection areas of each detection module corresponding to the region of interest overlap in the region of interest;

[0147] The control module 1508 is used to control each of the detection modules to be activated within the detection area corresponding to each of the detection modules.

[0148] The mechanical lidar control device provided in this application determines the region of interest (ROI) based on historical point cloud data. Then, based on the number of detection modules within the ROI, the detection area of ​​each detection module is set, enabling point cloud encryption within the ROI. By setting multiple detection modules, this application embodiment allows for real-time dynamic adjustment of the ROI during lidar operation, based on the determined ROI, thereby improving the lidar's energy utilization rate.

[0149] In one embodiment, the second determining module 1504 is further configured to:

[0150] Acquire at least one historical radar detection frame and determine the historical region of interest corresponding to each historical radar detection frame;

[0151] The region of interest is matched with each of the historical regions of interest to obtain the target historical regions of interest that match the region of interest, and the number of historical detection modules corresponding to the target historical regions of interest is determined.

[0152] Based on the historical number of detection modules and the preset strategy, the number of detection modules for the region of interest is set.

[0153] In one embodiment, the setting module 1504 is further configured to:

[0154] Determine the target segmentation region corresponding to each of the regions of interest in each segmentation region. The segmentation region is obtained by dividing the detection range of the lidar, and there is no overlap between the segmentation regions.

[0155] Based on the relative positions between the target segmentation regions, the detection area and phase of each detection module are set respectively.

[0156] In one embodiment, when both the number of detection modules and the number of regions of interest are 2, the segmented regions include 4, and each segmented region has a field of view of 90°; the setting module 1504 is further configured to:

[0157] When the two regions of interest are respectively included in the two target segmentation regions and the field of view of each region of interest does not exceed 90°, the detection area size of the two detection modules is set between 180° and 360°, and the phase of the two detection modules is set so that the phase difference between the two detection modules is between 90° and 180°.

[0158] In one embodiment, when the two target segmentation regions are symmetrical about the radar center, the two boundaries of one region of interest are the start position of the detection region of one detection module and the end position of the detection region of the other detection module, respectively.

[0159] The two boundaries of the other region of interest are the start position of the detection region of one detection module and the end position of the detection region of the other detection module, respectively.

[0160] The detection area of ​​both detection modules is greater than 180°.

[0161] In one embodiment, when the number of detection modules and regions of interest are both 2, the segmented regions include 4, and the field of view of each segmented region is 90°; the setting module 1504 is further configured to:

[0162] When the field of view of at least one of the regions of interest is greater than 90°, the size of the detection area of ​​at least one of the detection modules is set between 270° and 360°.

[0163] The modules in the aforementioned mechanical lidar control device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in the processor of a computer device in hardware form or independent of it, or stored in the memory of a computer device in software form, so that the processor can call and execute the operations corresponding to each module.

[0164] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 16 As shown, the computer device includes a processor, memory, and a network interface connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The network interface is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements a mechanical lidar control method.

[0165] Those skilled in the art will understand that Figure 16 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0166] In one embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method embodiments.

[0167] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon that, when executed by a processor, implements the steps in the above method embodiments.

[0168] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above method embodiments.

[0169] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties.

[0170] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0171] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0172] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A mechanical lidar control method, characterized in that, The method includes: Based on historical point cloud data, identify at least one region of interest; For any of the regions of interest, determine the number of detection modules corresponding to the region of interest; Based on the number of detection modules corresponding to each region of interest, the detection area of ​​each detection module is set respectively. For any region of interest, the detection areas of each detection module corresponding to the region of interest overlap in the region of interest. Each of the aforementioned detection modules is controlled to activate within its corresponding detection area; Determining the number of detection modules corresponding to any given region of interest includes: Acquire multiple historical radar detection frames and the historical point cloud data corresponding to each historical radar detection frame; For any of the historical radar detection frames, detect sparse points in the historical point cloud data corresponding to the historical radar detection frame. For any target sparse point, each historical radar detection frame and the target sparse point are matched according to the sparse points corresponding to each historical radar detection frame. If a preset number of historical radar detection frames are matched with the target sparse point, at least one region of interest is constructed according to the relative position of the target sparse point and the lidar, wherein the target sparse point is any of the sparse points.

2. The method according to claim 1, characterized in that, Determining the number of detection modules corresponding to any given region of interest includes: Acquire at least one historical radar detection frame and determine the historical region of interest corresponding to each historical radar detection frame; The region of interest is matched with each of the historical regions of interest to obtain the target historical regions of interest that match the region of interest, and the number of historical detection modules corresponding to the target historical regions of interest is determined. Based on the historical number of detection modules and the preset strategy, the number of detection modules for the region of interest is set.

3. The method according to claim 1, characterized in that, The step of setting the detection area of ​​each detection module according to the number of detection modules corresponding to each region of interest includes: Determine the target segmentation region corresponding to each of the regions of interest in each segmentation region. The segmentation region is obtained by dividing the detection range of the lidar, and there is no overlap between the segmentation regions. Based on the relative positions between the target segmentation regions, the detection area and phase of each detection module are set respectively.

4. The method according to claim 3, characterized in that, When both the number of detection modules and the number of regions of interest are two, the segmented regions include four, and each segmented region has a field of view of 90°; the step of setting the detection area and phase of each detection module according to the relative position between each target segmented region includes: When the two regions of interest are respectively included in the two target segmentation regions and the field of view of each region of interest does not exceed 90°, the detection area size of the two detection modules is set between 180° and 360°, and the phase of the two detection modules is set so that the phase difference between the two detection modules is between 90° and 180°.

5. The method according to claim 4, characterized in that, In the case where the two target segmentation regions are symmetrical about the radar center, the two boundaries of one region of interest are the start position of the detection region of one detection module and the end position of the detection region of the other detection module, respectively. The two boundaries of the other region of interest are the start position of the detection region of one detection module and the end position of the detection region of the other detection module, respectively. The detection area of ​​both detection modules is greater than 180°.

6. The method according to claim 3, characterized in that, With two detection modules and two regions of interest, the segmented regions comprise four, each with a field of view of 90°. The step of setting the detection region and phase of each detection module based on the relative positions of the target segmented regions includes: When the field of view of at least one of the regions of interest is greater than 90°, the size of the detection area of ​​at least one of the detection modules is set between 270° and 360°.

7. A mechanical lidar control device, characterized in that, The device includes: The first determination module is used to determine at least one region of interest based on historical point cloud data; The second determining module is used to determine the number of detection modules corresponding to any given region of interest. The setting module is used to set the detection area of ​​each detection module according to the number of detection modules corresponding to each region of interest. For any region of interest, the detection areas of each detection module corresponding to the region of interest overlap in the region of interest. A control module is used to control each of the detection modules to be activated within the detection area corresponding to each of the detection modules. The second determining module is specifically used to acquire multiple historical radar detection frames and historical point cloud data corresponding to each historical radar detection frame; for any historical radar detection frame, detect sparse points in the historical point cloud data corresponding to the historical radar detection frame; for any target sparse point, match each historical radar detection frame and the target sparse point according to the sparse points corresponding to each historical radar detection frame, and when a preset number of historical radar detection frames match the target sparse point, construct at least one region of interest according to the relative position of the target sparse point and the lidar, wherein the target sparse point is any of the sparse points.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

Images