A region-triggered perception method for a laser radar and related device

Abstract

The application provides a regional trigger perception method of a laser radar and related equipment, comprising: after obtaining point cloud frame data output by the laser radar, preprocessing is performed to extract a spatial point set and perform plane fitting, a spatial reference datum is obtained, and a calibration space coordinate system is established; the spatial point set is mapped to the calibration space coordinate system to obtain a three-dimensional point set and a two-dimensional projection point set in the calibration coordinate system; based on the three-dimensional point set and the two-dimensional projection point set, a plurality of trigger regions are generated in a plane corresponding to the spatial reference datum according to a preset direction division rule and a distance layering rule; and point cloud data in each trigger region is detected, when the point cloud in any trigger region meets a preset trigger condition, it is determined that the corresponding trigger region is in an activated state, and corresponding radar prompt information is generated. Through the above scheme, the spatial consistency and response accuracy of the detection result are improved, and the problem of inconsistency between regional division and actual space in a complex installation environment is overcome.

Description

Technical Field

[0001] This application relates to the field of point cloud data processing technology, and more specifically, to a region-triggered sensing method and related equipment for lidar. Background Technology

[0002] With the rapid development of autonomous driving, intelligent security and industrial automation, LiDAR, as a sensor capable of acquiring high-precision spatial distance information, is gradually becoming an important means of achieving safety early warning, area monitoring and behavior response in many application scenarios.

[0003] Currently, the common approach to triggering perception is to pre-define fixed spatial areas. This involves directly defining several regular areas (such as rectangular or sector-shaped areas) in the lidar coordinate system or the equipment installation coordinate system, and statistically analyzing the point cloud data entering these areas. When the number of point clouds within an area reaches a preset threshold or the distance meets a specific condition, the area is determined to be triggered, and corresponding prompts or control signals are output to achieve rapid detection of targets within a specific range and judgment of area intrusion.

[0004] Although target area triggering detection can be achieved by pre-setting a fixed area, when the lidar installation location is tilted, the ground is uneven, or the equipment deployment changes, the fixed area is difficult to keep in line with the actual spatial structure, which can easily lead to area division offset, thereby affecting the accuracy of triggering detection and reducing the system's perception reliability. Summary of the Invention

[0005] The embodiments of this application provide a region-triggered sensing method and related equipment for lidar, which can overcome the problem of inconsistency between region division and actual space in complex installation environments.

[0006] Other features and advantages of this application will become apparent from the following detailed description, or may be learned in part from practice of this application.

[0007] According to one aspect of the embodiments of this application, a method for triggering perception of a lidar area is provided, comprising: acquiring point cloud frame data output by a lidar; preprocessing the point cloud frame data to extract a spatial point set; performing plane fitting based on ground points in the spatial point set to obtain a spatial reference datum, and establishing a calibration spatial coordinate system based on the spatial reference datum; mapping the spatial point set to the calibration spatial coordinate system to obtain a three-dimensional point set in the calibration coordinate system; extracting the coordinate components of each point in the three-dimensional point set in the X-axis and Y-axis directions in the calibration spatial coordinate system to construct a two-dimensional projection point set; generating multiple trigger regions in the plane corresponding to the spatial reference datum according to preset direction division rules and distance layering rules based on the three-dimensional point set and the two-dimensional projection point set; detecting the point cloud data in each of the trigger regions; and determining that the corresponding trigger region is in an active state when the point cloud in any trigger region meets a preset trigger condition, and generating corresponding radar prompt information.

[0008] In some embodiments of this application, based on the aforementioned scheme, the step of acquiring point cloud frame data output by the lidar and preprocessing the point cloud frame data to extract a spatial point set includes: acquiring point cloud frame data from the lidar at a preset frame rate; performing distance filtering on the point cloud frame data to obtain a distance-filtered point cloud; performing intensity filtering on the distance-filtered point cloud to obtain an intensity-filtered point cloud; performing voxel downsampling on the intensity-filtered point cloud to obtain a downsampled point cloud; performing statistical outlier filtering on the downsampled point cloud, calculating the average distance between each point and a preset number of its nearest neighbors, and removing points whose average distance exceeds a preset multiple of the global mean standard deviation as outliers to obtain a spatial point set.

[0009] In some embodiments of this application, based on the aforementioned scheme, the step of performing plane fitting based on the ground points in the spatial point set to obtain a spatial reference datum, and establishing a calibration spatial coordinate system based on the spatial reference datum, includes: performing height threshold screening on the spatial point set, extracting points with height values ​​lower than a preset height threshold as candidate ground point sets; performing plane fitting on the candidate ground point sets, randomly selecting three non-collinear candidate ground points in each iteration to construct a candidate plane model, calculating the vertical distance from each point in the candidate ground point set to the candidate plane model; defining points with vertical distances less than a preset interior point distance threshold as interior points, counting the number of interior points, and using the candidate plane model with the most interior points as the fitting result to obtain the spatial reference datum; and establishing a calibration spatial coordinate system through the spatial reference datum and the original coordinate system of the lidar.

[0010] In some embodiments of this application, based on the foregoing scheme, the step of establishing a calibration spatial coordinate system using the spatial reference datum and the original coordinate system of the lidar includes: using the plane normal vector of the spatial reference datum as the Z-axis direction of the calibration space; using the projection direction of the lidar mounting orientation onto the spatial reference datum as the X-axis direction of the calibration space; determining the Y-axis direction of the calibration space based on the Z-axis direction and the X-axis direction; using the perpendicular point obtained by projecting the origin of the lidar's original coordinate system along the Z-axis direction onto the spatial reference datum as the origin of the coordinate system; and establishing a calibration spatial coordinate system using the Z-axis direction, the X-axis direction, the Y-axis direction, and the origin of the coordinate system.

[0011] In some embodiments of this application, based on the foregoing scheme, the step of mapping the spatial point set to the calibration spatial coordinate system to obtain a three-dimensional point set in the calibration coordinate system, and extracting the coordinate components of each point in the three-dimensional point set in the X-axis and Y-axis directions in the calibration spatial coordinate system to construct a two-dimensional projection point set includes: constructing a rotation matrix and a translation vector from the original coordinate system of the lidar to the calibration spatial coordinate system based on the origin of the calibration spatial coordinate system and the direction vectors of the X-axis, Y-axis, and Z-axis; sequentially transforming the coordinates of each point in the spatial point set according to the rotation matrix and the translation vector to obtain a three-dimensional point set in the calibration coordinate system; and extracting the coordinate components of each point in the three-dimensional point set in the X-axis and Y-axis directions in the calibration spatial coordinate system to construct a two-dimensional projection point set using the coordinate components of the X-axis and Y-axis.

[0012] In some embodiments of this application, based on the foregoing scheme, the step of generating multiple trigger regions in the plane corresponding to the spatial reference datum according to preset direction division rules and distance layering rules based on the three-dimensional point set and the two-dimensional projection point set includes: calculating the covariance matrix of the two-dimensional projection point set, and performing eigenvalue decomposition on the covariance matrix to obtain several eigenvalues; taking the eigenvector corresponding to the maximum value among the eigenvalues ​​as the principal direction, and taking the eigenvector corresponding to the minimum value among the eigenvalues ​​orthogonal to the principal direction as the secondary direction; taking the origin of the calibration spatial coordinate system as the base point, along the principal direction... The plane corresponding to the spatial reference datum is divided into multiple distance level intervals according to multiple preset distance level thresholds; the maximum and minimum values ​​of the three-dimensional point set in the Z-axis direction within each distance level interval are extracted as the upper and lower boundaries; the left and right boundaries of the trigger area in the secondary direction are determined based on the distribution range of the two-dimensional projection point set in the secondary direction within each distance level interval; the starting distance threshold and ending distance threshold along the main direction corresponding to each distance level interval are used as the front and back boundaries of the corresponding trigger area in the main direction; and the corresponding trigger area is generated based on the upper and lower boundaries, left and right boundaries, and front and back boundaries of each distance level interval.

[0013] In some embodiments of this application, based on the aforementioned scheme, the step of detecting point cloud data within each of the trigger areas, determining that the corresponding trigger area is in an active state when the point cloud in any trigger area meets a preset trigger condition, and generating corresponding radar prompt information includes: extracting point cloud data whose spatial coordinates fall within the range of each of the trigger areas from the three-dimensional point set; counting the number of points in the point cloud data, and calculating the difference between the maximum and minimum values ​​of the point cloud data in the height direction as the height span value; determining whether the number of points exceeds a preset point cloud number threshold and whether the height span value exceeds a preset height span threshold, and when both are met, determining that the corresponding trigger area is in an active state; generating radar prompt information based on the position information of the active trigger area in the calibration spatial coordinate system and the corresponding distance level identifier.

[0014] According to another aspect of the embodiments of this application, a region-triggered perception system for lidar is provided, comprising: a data acquisition module, configured to acquire point cloud frame data output by lidar, preprocess the point cloud frame data to extract a spatial point set; a plane fitting module, configured to perform plane fitting based on ground points in the spatial point set to obtain a spatial reference datum, and establish a calibration spatial coordinate system based on the spatial reference datum; a point set mapping module, configured to map the spatial point set to the calibration spatial coordinate system to obtain a three-dimensional point set in the calibration coordinate system, and extract the coordinate components of each point in the three-dimensional point set in the X-axis and Y-axis directions in the calibration spatial coordinate system to construct a two-dimensional projection point set; a region generation module, configured to generate multiple trigger regions in the plane corresponding to the spatial reference datum according to preset direction division rules and distance layering rules based on the three-dimensional point set and the two-dimensional projection point set; and a point cloud detection module, configured to detect the point cloud data in each of the trigger regions, and when the point cloud in any trigger region meets a preset trigger condition, determine that the corresponding trigger region is in an active state and generate corresponding radar prompt information.

[0015] According to another aspect of the embodiments of this application, an electronic device is provided, including a memory and a processor, the memory storing a computer program executable on the processor, the processor executing the computer program to implement the area-triggered sensing method of the lidar described in any one of the above.

[0016] According to another aspect of the embodiments of this application, a computer-readable storage medium is provided having a computer program stored thereon, which, when run by a processor, causes the processor to perform the area-triggered sensing method of a lidar as described in any one of the above embodiments.

[0017] Compared with existing technologies, this application has the following advantages: By preprocessing the point cloud frame data output by the lidar, a stable and reliable spatial point set is obtained. Based on the ground points extracted from the spatial point set, plane fitting is performed to construct a spatial reference benchmark, and a calibration spatial coordinate system is established in conjunction with the lidar installation orientation. The spatial point set is then mapped to the calibration spatial coordinate system to obtain a three-dimensional point set, and the X-axis and Y-axis coordinates are extracted to construct a two-dimensional projection point set. By performing principal direction analysis on the two-dimensional projection point set and combining it with distance layering thresholds, the plane corresponding to the spatial reference benchmark is partitioned, and the upper and lower boundaries of each triggering area are determined in conjunction with the height distribution of the three-dimensional point set, forming a set of triggering areas with a clear spatial range. Furthermore, by thresholding the number of point clouds and the height span within each triggering area, activation recognition of the triggering area is achieved, thereby realizing stable perception and partitioned triggering of the target area, improving the spatial consistency and response accuracy of the detection results, and overcoming the problem of inconsistency between area division and actual space in complex installation environments. Attached Figure Description

[0018] Figure 1 This is a flowchart illustrating a region-triggered sensing method for lidar provided in an embodiment of the present invention; Figure 2 This is a schematic diagram of the point cloud preprocessing process provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of the planar fitting effect of the RANSAC algorithm provided in the embodiment of the present invention; Figure 4 This is a schematic diagram of the spatial reference plane provided in an embodiment of the present invention; Figure 5 This is a schematic diagram of the process for constructing a calibration spatial coordinate system provided in an embodiment of the present invention; Figure 6 This is a top view of the spatial division of the trigger area provided in an embodiment of the present invention; Figure 7 This is a schematic block diagram of the structure of a lidar area-triggered sensing system provided in an embodiment of the present invention. Detailed Implementation

[0019] Exemplary embodiments will now be described in a more comprehensive manner with reference to the accompanying drawings. However, the exemplary embodiments can be implemented in various forms and should not be construed as limited to these examples; rather, these embodiments are provided so that this application will be more comprehensive and complete, and will fully convey the concept of the exemplary embodiments to those skilled in the art.

[0020] Furthermore, the features, structures, or characteristics described in this application can be combined in any suitable manner in one or more embodiments. Numerous specific details are provided in the following description to provide a full understanding of the embodiments of this application. However, those skilled in the art will recognize that when implementing the technical solutions of this application, not all the detailed features in the embodiments may be used, one or more specific details may be omitted, or other methods, elements, devices, steps, etc., may be employed.

[0021] The block diagrams shown in the accompanying drawings are merely functional entities and do not necessarily correspond to physically independent entities. That is, these functional entities can be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor devices and / or microcontroller devices.

[0022] The flowcharts shown in the accompanying drawings are merely illustrative and do not necessarily include all content and operations / steps, nor do they necessarily need to be performed in the described order. For example, some operations / steps can be broken down, while others can be combined or partially combined; therefore, the actual execution order may change depending on the specific circumstances.

[0023] It should be noted that "multiple" in this article refers to two or more. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, or B alone. The character " / " generally indicates that the preceding and following related objects have an "or" relationship.

[0024] The technical solution of the present invention will be further described below with reference to the accompanying drawings and specific embodiments.

[0025] like Figure 1 As shown in the example, this application provides a region-triggered perception method for LiDAR, which can be used to structurally divide the spatial region within the LiDAR's perception range and to achieve region-triggered detection and alerting based on point cloud distribution. The method specifically includes the following steps: Step S100: Obtain the point cloud frame data output by the lidar, and preprocess the point cloud frame data to extract the spatial point set.

[0026] The lidar outputs continuous point cloud frame data according to a preset scanning mode. Each point cloud frame consists of multiple spatial points, and each spatial point includes at least three-dimensional coordinate information (x, y, z) and reflection intensity information. The point cloud frame data is preprocessed to remove outliers, noise points, and invalid points, thereby obtaining a spatial point set that reflects the actual spatial structure. The preprocessing process includes multi-level filtering and downsampling of the point cloud frame data to reduce data redundancy and improve the stability of the point cloud distribution. The execution order of each level of processing is: distance filtering, intensity filtering, spatial downsampling, and outlier removal.

[0027] For example, a frame of point cloud data contains about 120,000 spatial points, some of which are located in distant areas beyond the actual detection range, or have low intensity values ​​due to insufficient reflection. After preprocessing, the number of points is compressed to about 30,000 to 50,000, while keeping the spatial structure features unchanged, thus obtaining a spatial point set.

[0028] like Figure 2 As shown, in another example, step S100 can also be implemented in the following manner: Point cloud frame data is acquired from the LiDAR at a preset frame rate. The point cloud frame data is then filtered by distance to remove points whose distance values ​​exceed the preset effective detection range, thus obtaining a distance-filtered point cloud.

[0029] The preset frame rate is set based on the hardware output capability of the LiDAR and the real-time requirements of the application scenario, with a value ranging from 5Hz to 20Hz. Distance filtering involves calculating the Euclidean distance from each spatial point to the LiDAR origin and comparing this distance value with the preset effective detection range to eliminate invalid points at long distances.

[0030] The Euclidean distance d is calculated as follows: d=√(x 2 +y 2 +z 2 ); The preset effective detection range includes a minimum distance threshold d_min and a maximum distance threshold d_max. d_min is used to remove noise points close to the radar, and d_max is used to remove distant points that are outside the effective detection range.

[0031] For example, in an indoor security scenario, d_min can be set to 0.2m and d_max can be set to 20m. When the distance d of a certain spatial point is less than 0.2m or greater than 20m, the spatial point is removed to obtain a distance-filtered point cloud.

[0032] Intensity filtering is applied to the distance-filtered point cloud to remove points with reflection intensity below a preset intensity threshold, resulting in an intensity-filtered point cloud. Voxel downsampling is then performed on the intensity-filtered point cloud. The point cloud space is divided into grids with a preset voxel size, and all points within each grid are replaced with the centroid coordinates of all points within that grid, resulting in a downsampled point cloud.

[0033] Intensity filtering is based on the reflection intensity value returned by the lidar. By setting a preset intensity threshold I_th, points with reflection intensity lower than the threshold are removed to reduce unstable points caused by material light absorption or long-distance attenuation.

[0034] Voxel downsampling employs a voxel grid filter algorithm, which reduces point cloud density by dividing the space into fixed-size three-dimensional grids (voxel units) and replacing all points in a voxel with the centroid coordinates of all points within that voxel.

[0035] The preset voxel size is set based on the spatial resolution requirements of the application scenario, and its value ranges from 0.05m to 0.2m.

[0036] The centroid coordinates are calculated as follows: x_c = (1 / n)∑x_i; y_c=(1 / n)∑y_i; z_c = (1 / n)∑z_i; Where n is the number of points within the voxel.

[0037] For example, in an industrial site monitoring scenario, the preset intensity threshold can be set to I_th=15 (based on the standard intensity range of 0~255 for lidar), and the voxel size can be set to 0.1m×0.1m×0.1m, compressing the original approximately 50,000 points to approximately 15,000 points, thereby obtaining a downsampled point cloud.

[0038] Statistical outlier filtering is performed on the downsampled point cloud. The average distance between each point and its preset number of nearest neighbors is calculated. Points whose average distance exceeds the preset multiple standard deviation of the global mean are removed as outliers to obtain a spatial point set.

[0039] The statistical outlier filtering uses the Statistical Outlier Removal algorithm, which detects anomalies based on the distance distribution between a point and its neighboring points.

[0040] The specific implementation method is as follows: For each point in the downsampled point cloud, search for its k nearest neighbors (k is the preset number of neighboring points), calculate the average distance d_avg from the point to the k neighboring points; construct a distance distribution for the average distance d_avg of all points, and calculate the global mean μ and standard deviation σ.

[0041] Points that meet the following conditions are identified as outliers and removed: d_avg>μ+α·σ; Where α is a preset multiplier coefficient.

[0042] The value of k ranges from 10 to 50, and the value of α ranges from 1.0 to 2.5.

[0043] For example, in a scenario with relatively uniform point cloud density, k=20 and α=1.5 can be set. When the average distance d_avg of a certain spatial point is significantly greater than the overall distribution, it is judged as an isolated noise point and removed. Finally, a spatial point set is obtained, with the number of points stable between about 10,000 and 20,000, and it has continuous spatial structure distribution characteristics.

[0044] Step S200: Perform plane fitting based on the ground points in the spatial point set to obtain a spatial reference datum, and establish a calibration spatial coordinate system based on the spatial reference datum.

[0045] A set of points representing the actual ground structure is identified from the spatial point set. This point set is then fitted with a plane to obtain a spatial reference plane, which is used as a unified spatial reference to construct a calibration spatial coordinate system. The spatial reference plane is used to eliminate the influence of the lidar installation attitude on the point cloud distribution, ensuring that region division is based on the actual ground orientation.

[0046] For example, when the lidar is installed at an angle, the ground points in the original point cloud are distributed at an angle in the original coordinate system. The spatial reference benchmark obtained by plane fitting can restore the true horizontal state of the ground, so that the area division is consistent with the actual spatial structure.

[0047] In another example, step S200 can also be implemented in the following way: The spatial point set is filtered by height threshold, and points with height values ​​lower than the preset height threshold are extracted as candidate ground point sets.

[0048] The height value is the coordinate value of a point along the Z-axis in the original coordinate system of the lidar. By setting a preset height threshold H_th, the spatial point set is initially filtered, and points with height values ​​lower than the threshold are selected as candidate ground point sets.

[0049] The preset height threshold is set based on the installation height of the lidar and the range of ground undulations, and its value ranges from 10% to 50% of the installation height of the lidar.

[0050] For example, when the lidar is installed at a height of 1.5m, the height threshold can be set to H_th=0.5m, and points with Z≤0.5m can be extracted as a candidate ground point set, thereby excluding most target points that are higher than the ground.

[0051] The random sampling consensus algorithm is used to fit the candidate ground point set to the plane. In each iteration, three non-collinear candidate ground points are randomly selected to construct a candidate plane model, and the vertical distance from each point in the candidate ground point set to the candidate plane model is calculated.

[0052] like Figure 3 As shown, the random sample consensus algorithm is the RANSAC (Random Sample Consensus) algorithm, which achieves robust fitting of noisy data through multiple random sampling and model evaluation.

[0053] In each iteration, three non-collinear points P1(x1,y1,z1), P2(x2,y2,z2), and P3(x3,y3,z3) are randomly selected from the candidate ground point set, and the plane equation is constructed as follows: Ax + By + Cz + D = 0; The normal vector n = (A, B, C) is obtained by the cross product of vectors (P2-P1) and (P3-P1).

[0054] For any point Pi(xi,yi,zi) in the candidate ground point set, its perpendicular distance di to the plane is calculated as follows: di = |Axi + Byi + Czi + D| / √(A 2 +B 2 +C 2 ); For example, after randomly selecting three points to construct a candidate plane in one iteration, the distance values ​​of approximately 8,000 candidate ground points are calculated one by one to evaluate how well the plane model fits the ground points.

[0055] like Figure 4 and Figure 5 As shown, points with a vertical distance less than a preset interior point distance threshold are defined as interior points. The number of interior points is counted, and the candidate plane model with the most interior points is used as the fitting result to obtain the spatial reference benchmark.

[0056] The preset interior point distance threshold d_in is used to determine whether a point belongs to the current plane model, and its value ranges from 0.01m to 0.05m.

[0057] For each candidate planar model, points satisfying di ≤ d_in are designated as interior points, and the number of interior points is counted. After a predetermined number of iterations, the planar model with the largest number of interior points is selected as the final fitting result.

[0058] The number of iterations is set based on the point cloud size, and its value ranges from 50 to 200.

[0059] For example, if the interior point distance threshold is 0.05m, after 100 iterations, the number of interior points of a certain candidate plane model reaches 6500, which is the maximum value among all candidate models. Then, this plane model is used as the spatial reference benchmark.

[0060] A calibration spatial coordinate system is established by using a spatial reference datum and the original coordinate system of the lidar.

[0061] Based on the spatial reference benchmark and combined with the installation direction information of the lidar, the original coordinate system is reconstructed to obtain the calibration spatial coordinate system, which is consistent with the actual ground direction.

[0062] For example, after this step, the coordinate system deflection caused by the tilted installation is eliminated, ensuring that the Z-axis is always perpendicular to the ground.

[0063] The specific steps for establishing a calibration spatial coordinate system using a spatial reference datum and the original coordinate system of the lidar are as follows: The plane normal vector of the spatial reference datum is used as the Z-axis direction of the calibration space.

[0064] The normal vector n=(A,B,C) obtained from plane fitting is normalized: n_z = n / |n|; Where |n| is the magnitude of the normal vector. The normalized vector serves as the Z-axis direction of the calibration space coordinate system, thus ensuring that the Z-axis is perpendicular to the space reference datum.

[0065] For example, when the normal vector is (0.1, 0.2, 0.97), it is normalized and used as the Z-axis direction, so that it points to the direction perpendicular to the ground.

[0066] The X-axis direction of the calibration space is taken as the projection direction of the lidar installation orientation onto the spatial reference datum.

[0067] The lidar's mounting orientation is represented by the forward unit vector f=(fx,fy,fz) in the lidar's original coordinate system. Projecting this vector onto the spatial reference plane yields the X-axis direction vector: f_proj=f-(f·n_z)n_z; Then, f_proj is normalized to obtain the unit vector n_x.

[0068] For example, when the forward direction of the lidar is (1,0,0) and the ground normal is (0,0,1), its projection result is still (1,0,0), which is used as the X-axis direction.

[0069] The Y-axis direction of the calibration space is determined based on the Z-axis and X-axis directions according to the right-hand rule.

[0070] Calculate the Y-axis direction using the cross product of vectors: n_y=n_z×n_x Then, n_y is normalized to ensure that the X-axis, Y-axis, and Z-axis are orthogonal to each other, forming a right-handed coordinate system.

[0071] For example, when the Z-axis is (0,0,1) and the X-axis is (1,0,0), the Y-axis is (0,1,0) obtained by cross product.

[0072] The origin of the coordinate system is obtained by projecting the origin of the original coordinate system of the lidar along the Z-axis to the perpendicular point of the spatial reference datum. The origin of the coordinate system is located in the plane of the spatial reference datum and is consistent with the X-axis direction.

[0073] Project the origin O(0,0,0) of the original coordinate system along the Z-axis onto the spatial reference plane to obtain the perpendicular point O', which serves as the origin of the calibration spatial coordinate system.

[0074] The projection is calculated using the normalized normal vector n_z, specifically implemented as follows: t = -(n_z·O_vec+d_n; Where n_z is the normalized plane normal vector in step S200, O_vec = (0,0,0) is the origin of the original coordinate system, and d_n is the intercept term of the plane equation expressed in terms of the normalized normal vector, which is calculated as follows: d_n=-(n_zx·x_ref+n_zy·y_ref+n_zz·z_ref); Where (x_ref, y_ref, z_ref) are the coordinates of any known point on the spatial reference plane. The coordinates of the perpendicular point are calculated as follows: O'=O_vec+t·n_z=(t·n_zx, t·n_zy, t·n_zz); The above calculations use the normalized normal vector n_z throughout, which is consistent with the normalization process of the normal vector mentioned above, and avoids deviations in the projection results caused by using the unnormalized original normal vector.

[0075] Consistency constraints refer to the selection of the origin without changing the definition of the X-axis direction, thus stabilizing the coordinate system direction.

[0076] For example, when the normalized normal vector of the spatial reference plane is n_z = (0,0,1) and the plane intercept d_n = -1, t = 1, and the perpendicular point O' = (0,0,1) is used as the new coordinate origin.

[0077] Establish a calibration spatial coordinate system using the Z-axis, X-axis, Y-axis, and origin.

[0078] Using the three unit direction vectors n_x, n_y, and n_z obtained above, and the origin O', a calibration space coordinate system is constructed, and a coordinate transformation matrix can be further formed: R=[n_x;n_y;n_z]; T=O'; This enables coordinate transformation from the original coordinate system to the calibration space coordinate system.

[0079] For example, after the above construction is completed, the coordinates of any point in the new coordinate system can be calculated through matrix transformation, realizing a unified expression of point cloud data in standard space.

[0080] Step S300: Map the spatial point set to the calibration spatial coordinate system to obtain a three-dimensional point set in the calibration coordinate system. Extract the coordinate components of each point in the three-dimensional point set in the X-axis and Y-axis directions in the calibration spatial coordinate system to construct a two-dimensional projection point set.

[0081] Based on the calibration spatial coordinate system established in step S200, each point in the spatial point set is transformed from the original coordinate system of the lidar to the calibration spatial coordinate system to eliminate the influence of the installation attitude on the point cloud distribution. After the transformation, the Z-axis direction of the three-dimensional point set represents the height information relative to the spatial reference datum, while the X-axis and Y-axis directions represent the spatial distribution in the plane.

[0082] Based on this, the three-dimensional point set is reduced in dimensionality, retaining only the coordinate components in the X and Y axes, thereby constructing a two-dimensional projection point set.

[0083] For example, if the coordinates of a point in the original point cloud are (1.2, 0.5, 0.3), after coordinate transformation, the coordinates in the calibration space coordinate system are (1.0, 0.6, 0.05), and the corresponding two-dimensional projection point is (1.0, 0.6).

[0084] In another example, step S300 can also be implemented in the following way: Based on the origin of the calibration spatial coordinate system and the direction vectors of the X, Y, and Z axes, construct the rotation matrix and translation vector from the original coordinate system of the lidar to the calibration spatial coordinate system.

[0085] Construct the rotation matrix R using the unit direction vectors n_x, n_y, and n_z obtained in step S200: R=[n_xx n_xy n_xz, n_yx n_yy n_yz, n_zx n_zy n_zz]; The translation vector T is the coordinate of the origin O' of the calibration spatial coordinate system in the original coordinate system.

[0086] The coordinate transformation relationship is: P' = R·(PT); Where P is the original coordinate and P' is the calibration coordinate.

[0087] For example, when n_x=(1,0,0), n_y=(0,1,0), n_z=(0,0,1), and T=(0,0,1), it means that the overall coordinate system is translated 1m along the Z-axis.

[0088] The coordinates of each point in the spatial point set are transformed sequentially using the rotation matrix and translation vector to obtain the three-dimensional point set in the calibration coordinate system.

[0089] Perform the above coordinate transformation on each point in the spatial point set one by one, and transform all points into a unified calibration spatial coordinate system, thereby obtaining a three-dimensional point set.

[0090] The calculation process can be implemented point by point through matrix multiplication, or through batch matrix operations.

[0091] For example, for 10,000 spatial points, the coordinate transformation of all points can be completed simultaneously through a single matrix operation to obtain a three-dimensional point set.

[0092] Extract the coordinate components of each point in the three-dimensional point set along the X and Y axes in the calibration space coordinate system, discard the component along the Z axis, and construct a two-dimensional projection point set using the coordinate components along the X and Y axes.

[0093] Each three-dimensional point P'(x',y',z') is mapped to a two-dimensional point P''(x',y'), thus forming a two-dimensional projection point set.

[0094] This processing method represents the spatial point cloud on the plane corresponding to the spatial reference datum, so that the region division process is calculated only based on planar coordinates, thereby reducing computational complexity.

[0095] For example, when a three-dimensional point set contains 15,000 points, 15,000 two-dimensional points are generated to form a two-dimensional projection point set.

[0096] Step S400: Based on the three-dimensional point set and the two-dimensional projection point set, generate multiple trigger areas in the plane corresponding to the spatial reference datum according to preset direction division rules and distance layering rules.

[0097] Based on the two-dimensional projection point set, the distribution range in the plane is determined, and combined with the height information in the three-dimensional point set, multiple trigger regions with spatial ranges are divided in the plane corresponding to the spatial reference datum. Each trigger region is a three-dimensional spatial region with front, back, left, right and top and bottom boundaries.

[0098] For example, in a forward detection scenario, the space in front is divided into several strip-shaped regions, each corresponding to a different distance range.

[0099] like Figure 6 As shown, in another example, step S400 can also be implemented in the following manner: A principal direction analysis is performed on the two-dimensional projection point set to calculate the covariance matrix of the two-dimensional projection point set, and eigenvalue decomposition is performed on the covariance matrix to obtain several eigenvalues.

[0100] The covariance matrix is ​​calculated as follows: Cov=[Var(x)Cov(x,y) Cov(y,x)Var(y)]; Where Var(x) and Var(y) represent the variances in the X-axis and Y-axis directions, respectively.

[0101] The matrix is ​​decomposed into eigenvalues ​​to obtain two eigenvalues ​​λ1 and λ2 and their corresponding eigenvectors.

[0102] For example, when λ1 is much greater than λ2, it indicates that the point cloud is more concentrated along a certain direction.

[0103] The eigenvector corresponding to the maximum value among the eigenvalues ​​is taken as the principal direction, and the eigenvector corresponding to the minimum value among the eigenvalues ​​that is orthogonal to the principal direction is taken as the secondary direction.

[0104] The eigenvector corresponding to the largest eigenvalue is selected as the main direction n_main, and the other eigenvector is selected as the secondary direction n_sub, and the two are orthogonal.

[0105] For example, when the principal direction is (0.98, 0.2), the secondary direction is approximately (-0.2, 0.98).

[0106] Using the origin of the calibrated spatial coordinate system as the base point, the plane corresponding to the spatial reference datum is divided into multiple distance level intervals along the main direction according to multiple preset distance level thresholds.

[0107] The distance stratification threshold is a set of incremental values, such as {0m, 2m, 5m, 10m}, and adjacent thresholds constitute a distance stratification interval.

[0108] For example, interval 1 is 0~2m, interval 2 is 2~5m, and interval 3 is 5~10m.

[0109] Extract the maximum and minimum values ​​of the 3D point set in the Z-axis direction within each distance level interval as the upper and lower boundaries.

[0110] Points belonging to the same distance level interval are filtered, and their maximum and minimum Z-axis values ​​are calculated to determine the height range of the region.

[0111] For example, if the Z range is 0.0m to 1.8m within a certain interval, then the height boundary of this interval is [0.0, 1.8].

[0112] The left and right boundaries of the triggering area in the secondary direction are determined based on the distribution range of the two-dimensional projection point set in each corresponding distance level interval in the secondary direction.

[0113] Calculate the range of projection values ​​of all two-dimensional points in the corresponding interval in the second direction to obtain the minimum value L_min and the maximum value L_max, thereby determining the left and right boundaries.

[0114] For example, within a certain interval, the secondary direction range is from -1.5m to 1.5m.

[0115] The starting distance threshold and ending distance threshold along the main direction corresponding to each distance level interval are used as the front and rear boundaries of the corresponding triggering area in the main direction.

[0116] The starting threshold of each distance level interval is used as the front boundary, and the ending threshold is used as the back boundary, thereby determining the range of the region in the main direction.

[0117] For example, for the interval 2~5m, its front and rear boundaries are 2m and 5m.

[0118] The corresponding triggering areas are generated based on the upper and lower boundaries, left and right boundaries, and front and back boundaries of each distance level interval.

[0119] The boundaries of the above three directions are combined to form a three-dimensional cuboid region as the trigger region.

[0120] For example, assuming the primary and secondary directions have been determined through eigenvalue decomposition of the covariance matrix, the spatial extent of a certain triggering region can be represented as: Main directional range: 2m to 5m (corresponding to the front and rear boundaries of this distance level interval); Secondary directional range: -1.5m to 1.5m (corresponding to the distribution range of the two-dimensional projection point set in the secondary direction within the corresponding distance level interval); Height range: 0m to 1.8m (corresponding to the maximum and minimum values ​​of the 3D point set in the Z-axis direction within this distance level interval); The aforementioned primary and secondary directions are the actual principal axis directions calculated based on the point cloud distribution characteristics. They are not necessarily consistent with the X-axis direction of the calibrated spatial coordinate system. The boundary description of the triggering area is based on the primary and secondary directions, rather than directly corresponding to the X-axis and Y-axis directions.

[0121] Step S500: Detect the point cloud data in each trigger area. When the point cloud in any trigger area meets the preset trigger conditions, determine that the corresponding trigger area is in an active state and generate the corresponding radar prompt information.

[0122] Perform statistical analysis on the point cloud within each trigger area, and determine whether the area has been triggered based on the point cloud distribution.

[0123] For example, when a significant point cloud cluster appears in a certain area and the height changes significantly, it is determined that the area has been triggered.

[0124] In another example, step S500 can also be implemented in the following way: Point cloud data whose spatial coordinates fall within the range of each trigger area are extracted from the 3D point set.

[0125] Filtering points by range judgment: X_min≤x≤X_max; Y_min≤y≤Y_max; Z_min≤z≤Z_max; Points that meet the conditions belong to the corresponding triggering area.

[0126] For example, approximately 500 points can be selected from a region.

[0127] The number of points in the point cloud data is counted, and the difference between the maximum and minimum values ​​in the height direction of the point cloud data is calculated as the height span value.

[0128] The number of point clouds is denoted as N, and the height span ΔH is defined as: ΔH = Z_max - Z_min; For example, when Z_max=1.8m and Z_min=0.2m, then ΔH=1.6m.

[0129] Determine whether the number of point clouds exceeds a preset point cloud number threshold and whether the height span value exceeds a preset height span threshold. If both conditions are met, determine that the corresponding trigger area is in an active state.

[0130] The point cloud quantity threshold N_th ranges from 50 to 300, and the height span threshold H_th ranges from 0.5m to 1.5m.

[0131] For example, the triggering condition is met when N=200 and ΔH=1.2m.

[0132] Radar prompt information is generated based on the location information of the active trigger area in the calibration spatial coordinate system and the corresponding distance level identifier. The radar prompt information includes the area number, trigger timestamp, and azimuth interval information relative to the X-axis of the calibration spatial coordinate system.

[0133] The area number is numbered in order of distance hierarchy, the timestamp is the current frame acquisition time, and the azimuth interval is determined according to the position of the area in the main direction and the secondary direction.

[0134] For example, if the area number is R2, the timestamp is 10:01:25, and the orientation range is 2~5m ahead, then the corresponding prompt message will be output.

[0135] In this embodiment, a stable and reliable spatial point set is obtained by performing range filtering, intensity filtering, voxel downsampling, and statistical outlier filtering on the point cloud frame data output by the lidar. Based on the ground points extracted from the spatial point set, plane fitting is performed to construct a spatial reference benchmark. A calibration spatial coordinate system is established in conjunction with the lidar installation orientation to achieve unified constraints on the spatial attitude of the point cloud. Furthermore, the spatial point set is mapped to the calibration spatial coordinate system to obtain a three-dimensional point set, and the X-axis and Y-axis coordinates are extracted to construct a two-dimensional projection point set, thereby reducing the computational complexity of region partitioning. Based on this, the two-dimensional projection point set is subjected to primary... By analyzing the direction and combining it with distance layering thresholds, the plane corresponding to the spatial reference benchmark is partitioned. At the same time, the upper and lower boundaries of each triggering area are determined by combining the height distribution of the three-dimensional point set, forming a set of triggering areas with a clear spatial range. Finally, by applying threshold judgments to the number of point clouds and the height span of each triggering area, the activation recognition of the triggering area is realized, and radar prompt information containing area number, timestamp, and azimuth interval information is output. This achieves stable perception and partitioned triggering of the target area, improves the spatial consistency and response accuracy of the detection results, and overcomes the problem of inconsistency between area division and actual space in complex installation environments.

[0136] like Figure 7 As shown, this application also provides a lidar area-triggered sensing system 10, comprising: The data acquisition module 11 is used to acquire point cloud frame data output by the lidar and preprocess the point cloud frame data to extract spatial point sets.

[0137] The plane fitting module 12 is used to perform plane fitting based on ground points in the spatial point set, obtain a spatial reference datum, and establish a calibration spatial coordinate system based on the spatial reference datum.

[0138] The point set mapping module 13 is used to map the spatial point set to the calibration spatial coordinate system to obtain the three-dimensional point set in the calibration coordinate system, and extract the coordinate components of each point in the three-dimensional point set in the X-axis and Y-axis directions in the calibration spatial coordinate system to construct the two-dimensional projection point set.

[0139] The region generation module 14 is used to generate multiple trigger regions in the plane corresponding to the spatial reference datum according to preset direction division rules and distance layering rules based on the three-dimensional point set and the two-dimensional projection point set.

[0140] The point cloud detection module 15 is used to detect the point cloud data in each trigger area. When the point cloud in any trigger area meets the preset trigger conditions, it determines that the corresponding trigger area is in an active state and generates the corresponding radar prompt information.

[0141] In this example, the data acquisition module 11 performs range filtering, intensity filtering, voxel downsampling, and statistical outlier filtering on the point cloud frame data output by the lidar to obtain a stable and consistent spatial point set. The plane fitting module 12 performs plane fitting based on the ground points in the spatial point set using a random sampling consensus algorithm to construct a spatial reference benchmark. Combined with the lidar installation orientation, a calibration spatial coordinate system is established to achieve unified spatial attitude constraints on the point cloud data. The point set mapping module 13 transforms the spatial point set into the calibration spatial coordinate system based on a rotation matrix and translation vector, obtaining a three-dimensional point set. The X-axis and Y-axis coordinates are extracted to construct a two-dimensional projection point set for practical application. The system performs dimensionality reduction representation of spatial data; the region generation module 14 performs main direction analysis based on the two-dimensional projection point set, and divides the corresponding plane of the spatial reference benchmark by combining the preset distance layering threshold. At the same time, it integrates the distribution of the three-dimensional point set in the Z-axis direction to determine the upper and lower boundaries of each trigger region, thereby constructing a set of trigger regions with a clear spatial range; the point cloud detection module 15 performs threshold judgment on the number of point clouds and the height span of each trigger region, identifies the activation status of the trigger region, and outputs radar prompt information containing region number, trigger timestamp and azimuth interval information, thereby realizing structured region division and stable trigger recognition of point cloud data, and improving the consistency and response accuracy of region perception.

[0142] This application also provides an electronic device, including a memory and a processor. The memory stores a computer program that can run on the processor. When the processor executes the computer program, it implements the aforementioned area-triggered sensing method for lidar.

[0143] In this example, by setting up a memory and processor in the electronic device, and storing a computer program in the memory for performing point cloud data processing, spatial reference benchmark construction, coordinate system mapping, region division, and trigger determination, the processor, when executing the computer program, sequentially completes the preprocessing of point cloud frame data, ground point plane fitting, calibration of spatial coordinate system construction, 3D point set mapping and 2D projection point set generation, trigger region generation, and point cloud trigger detection processing. This transforms the raw point cloud data of the lidar into region trigger information with spatial structure constraints, and outputs radar prompt information containing region number, timestamp, and azimuth interval, enabling real-time perception and trigger determination of spatial regions, and improving the point cloud data processing capability and region recognition accuracy of the electronic device in complex environments.

[0144] This application also provides a computer-readable storage medium having a computer program stored thereon, which, when run by a processor, causes the processor to execute the area-triggered sensing method for LiDAR as described above.

[0145] In this example, by storing a computer program in a computer-readable storage medium for performing point cloud data preprocessing, spatial reference datum construction, calibration spatial coordinate system establishment, point cloud coordinate mapping, 2D projection generation, region division, and trigger detection, the processor, when calling the computer program, can process the point cloud frame data output by the LiDAR step by step according to the preset processing flow, converting the disordered point cloud data into structured region data based on the spatial reference datum, and identifying the trigger region status based on the joint judgment rule of point cloud quantity and height span, and outputting the corresponding radar prompt information, thereby realizing the software deployment of the LiDAR region trigger perception method and enhancing the reusability and implementation flexibility of the point cloud data processing flow.

[0146] It should be noted that although several modules or units of the system for executing actions have been mentioned in the detailed description above, this division is not mandatory. In fact, according to the embodiments of this application, the features and functions of two or more modules or units described above can be embodied in one module or unit. Conversely, the features and functions of one module or unit described above can be further divided and embodied by multiple modules or units.

[0147] From the above description of the embodiments, those skilled in the art will readily understand that the exemplary embodiments described herein can be implemented by software or by combining software with necessary hardware. Therefore, the technical solutions according to the embodiments of this application can be embodied in the form of a software product, which can be stored on a non-volatile storage medium (such as a CD). The method, which is contained in or on a ROM, USB flash drive, external hard drive, etc., includes several instructions to cause an electronic device (which may be a personal computer, server, touch terminal, or network device, etc.) to execute the method according to the embodiments of this application.

[0148] Other embodiments of this application will readily occur to those skilled in the art upon consideration of the specification and practice of the embodiments disclosed herein. This application is intended to cover any variations, uses, or adaptations of this application that follow the general principles of this application and include common knowledge or customary techniques in the art not disclosed herein.

[0149] It should be understood that this application is not limited to the precise structure described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this application is limited only by the appended claims.

Claims

1. A method for area-triggered sensing using lidar, characterized in that, include: Acquire point cloud frame data output by lidar, and preprocess the point cloud frame data to extract spatial point sets; Plane fitting is performed on the ground points in the spatial point set to obtain a spatial reference datum, and a calibration spatial coordinate system is established based on the spatial reference datum; The spatial point set is mapped to the calibration spatial coordinate system to obtain a three-dimensional point set in the calibration coordinate system. The coordinate components of each point in the three-dimensional point set in the calibration spatial coordinate system in the X-axis and Y-axis directions are extracted to construct a two-dimensional projection point set. Based on the three-dimensional point set and the two-dimensional projection point set, multiple trigger regions are generated in the plane corresponding to the spatial reference datum according to preset direction division rules and distance layering rules. The point cloud data in each of the aforementioned trigger areas is detected. When the point cloud in any trigger area meets the preset trigger conditions, the corresponding trigger area is determined to be in an active state, and corresponding radar prompt information is generated.

2. The area-triggered sensing method for lidar according to claim 1, characterized in that, The step of acquiring point cloud frame data output by the lidar and preprocessing the point cloud frame data to extract a spatial point set includes: Point cloud frame data is acquired from the lidar at a preset frame rate, and the point cloud frame data is subjected to distance filtering to obtain a distance-filtered point cloud. Intensity filtering is applied to the distance-filtered point cloud to obtain an intensity-filtered point cloud, and voxel downsampling is performed on the intensity-filtered point cloud to obtain a downsampled point cloud; Statistical outlier filtering is performed on the downsampled point cloud. The average distance between each point and its preset number of nearest neighbors is calculated. Points whose average distance exceeds the preset multiple standard deviation of the global mean are removed as outliers to obtain a spatial point set.

3. The area-triggered sensing method for lidar according to claim 1, characterized in that, The step of performing plane fitting based on the ground points in the spatial point set to obtain a spatial reference datum, and establishing a calibration spatial coordinate system based on the spatial reference datum, includes: The spatial point set is filtered by height threshold, and points with height values ​​lower than a preset height threshold are extracted as candidate ground point sets; Plane fitting is performed on the candidate ground point set. In each iteration, three non-collinear candidate ground points are randomly selected to construct a candidate plane model. The vertical distance from each point in the candidate ground point set to the candidate plane model is calculated. Points whose vertical distance is less than a preset interior point distance threshold are defined as interior points. The number of interior points is counted, and the candidate plane model with the most interior points is used as the fitting result to obtain a spatial reference benchmark. A calibration spatial coordinate system is established using the spatial reference datum and the original coordinate system of the lidar.

4. The area-triggered sensing method for lidar according to claim 3, characterized in that, The step of establishing a calibration spatial coordinate system using the spatial reference datum and the original coordinate system of the lidar includes: The plane normal vector of the spatial reference datum is used as the Z-axis direction of the calibration space; The X-axis direction of the calibration space is defined by the projection direction of the lidar installation onto the spatial reference datum. The Y-axis direction of the calibration space is determined based on the Z-axis direction and the X-axis direction; The origin of the coordinate system is obtained by projecting the origin of the original coordinate system of the lidar along the Z-axis direction onto the spatial reference datum. A calibration spatial coordinate system is established using the Z-axis direction, the X-axis direction, the Y-axis direction, and the origin.

5. The area-triggered sensing method for lidar according to claim 1, characterized in that, The step of mapping the spatial point set to the calibration spatial coordinate system to obtain a three-dimensional point set in the calibration coordinate system, and extracting the coordinate components of each point in the three-dimensional point set in the X-axis and Y-axis directions in the calibration spatial coordinate system to construct a two-dimensional projection point set includes: Based on the origin of the calibration spatial coordinate system and the direction vectors of the X, Y, and Z axes, construct the rotation matrix and translation vector from the original coordinate system of the lidar to the calibration spatial coordinate system. The coordinates of each point in the spatial point set are transformed sequentially according to the rotation matrix and the translation vector to obtain the three-dimensional point set in the calibration coordinate system. Extract the coordinate components of each point in the three-dimensional point set along the X and Y axes in the calibration space coordinate system, and construct a two-dimensional projection point set using the coordinate components along the X and Y axes.

6. The area-triggered sensing method for lidar according to claim 1, characterized in that, The step of generating multiple trigger regions in the plane corresponding to the spatial reference datum according to preset direction division rules and distance layering rules based on the three-dimensional point set and the two-dimensional projected point set includes: Calculate the covariance matrix of the two-dimensional projection point set, and perform eigenvalue decomposition on the covariance matrix to obtain several eigenvalues; The eigenvector corresponding to the maximum value among the eigenvalues ​​is taken as the principal direction, and the eigenvector corresponding to the minimum value among the eigenvalues ​​that is orthogonal to the principal direction is taken as the secondary direction; Using the origin of the calibrated spatial coordinate system as the base point, the plane corresponding to the spatial reference datum is divided into multiple distance level intervals along the main direction according to multiple preset distance level thresholds; Extract the maximum and minimum values ​​of the 3D point set in the Z-axis direction within each distance level interval as the upper and lower boundaries; The left and right boundaries of the triggering area in the secondary direction are determined based on the distribution range of the two-dimensional projection point set in each corresponding distance level interval in the secondary direction. The starting distance threshold and ending distance threshold along the main direction corresponding to each distance level interval are used as the front and rear boundaries of the corresponding triggering area in the main direction. The corresponding triggering areas are generated based on the upper and lower boundaries, left and right boundaries, and front and back boundaries of each distance level interval.

7. The area-triggered sensing method for lidar according to claim 1, characterized in that, The step of detecting point cloud data within each of the trigger areas, determining that the corresponding trigger area is active when the point cloud data within any trigger area meets a preset trigger condition, and generating corresponding radar prompt information includes: Extract point cloud data whose spatial coordinates fall within the range of each of the three-dimensional point sets; The number of points in the point cloud data is counted, and the difference between the maximum and minimum values ​​of the point cloud data in the height direction is calculated as the height span value. Determine whether the number of point clouds exceeds a preset point cloud number threshold and whether the height span value exceeds a preset height span threshold. When both conditions are met, determine that the corresponding trigger area is in an active state. Radar alert information is generated based on the position information of the active triggering area in the calibration space coordinate system and the corresponding distance level identifier.

8. A region-triggered sensing system for lidar, characterized in that, include: The data acquisition module is used to acquire point cloud frame data output by the lidar, and to preprocess the point cloud frame data to extract spatial point sets. The plane fitting module is used to perform plane fitting based on the ground points in the spatial point set to obtain a spatial reference datum, and to establish a calibration spatial coordinate system based on the spatial reference datum; The point set mapping module is used to map the spatial point set to the calibration spatial coordinate system to obtain a three-dimensional point set in the calibration coordinate system, and to extract the coordinate components of each point in the three-dimensional point set in the X-axis and Y-axis directions in the calibration spatial coordinate system to construct a two-dimensional projection point set. The region generation module is used to generate multiple trigger regions in the plane corresponding to the spatial reference datum according to preset direction division rules and distance layering rules, based on the three-dimensional point set and the two-dimensional projection point set. The point cloud detection module is used to detect the point cloud data in each of the trigger areas. When the point cloud in any trigger area meets the preset trigger conditions, it determines that the corresponding trigger area is in an active state and generates the corresponding radar prompt information.

9. An electronic device, characterized in that, The system includes a memory and a processor, the memory storing a computer program that can run on the processor, and the processor executing the computer program to implement the area-triggered sensing method of the lidar according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, It stores a computer program that, when executed by a processor, causes the processor to perform the area-triggered perception method of the lidar as described in any one of claims 1 to 7.

Images