WO2023185927A1 - Method, apparatus and device for determining layering of point cloud of lidar, and storage medium - Google Patents

Abstract

A method, apparatus and device for determining layering of a point cloud of a LiDAR, and a storage medium. The LiDAR comprises a first field of view and a second field of view, wherein there is an overlapping region between the first field of view and the second field of view. The method comprises: acquiring a first alternative point of a point cloud of a LiDAR, wherein the first alternative point is a point which is included in a first field of view and is located in an overlapping region (step 41); acquiring a second alternative point of the point cloud of the LiDAR, wherein the second alternative point is a point which is included in a second field of view and is located in the overlapping region (step 42); and according to a comparison result between a predetermined feature of the first alternative point and a predetermined feature of the second alternative point, determining whether the point cloud is layered (step 43).

Description

Methods, devices, equipment and storage media for determining lidar point cloud stratification

priority information

This application is based on the application with CN application number 202210328377.4 and the filing date is March 30, 2022, and claims its priority. The disclosure content of the CN application is hereby incorporated into this application as a whole.

Technical field

The present disclosure relates to the technical field of lidar, and in particular, to a method, device, equipment and storage medium for determining layering of lidar point clouds.

Background technique

LiDAR (Light Detection and Ranging) is the abbreviation of laser detection and ranging system. Generally, an infrared laser is used as the emitting light source to emit a laser beam in a certain direction around the LiDAR. When the laser beam encounters an object, diffuse reflection occurs, and part of the scattered light of the laser returns to the laser receiving system. The lidar information processing module can calculate the distance between the lidar and the object based on the speed of light based on the time interval between transmitting and receiving laser signals.

In a very short time, by emitting laser beams in multiple directions around the LiDAR and measuring the distance, a frame of 3D laser point cloud can be output. In various application fields where lidar is used as a key sensor for sensing the surrounding environment, obstacles can be sensed based on information such as the spatial position of point clouds. Therefore, LiDAR is widely used in fields such as autonomous driving, robot obstacle avoidance, vehicle-road collaboration in smart cities, and surveying and mapping.

In related technologies, if lidar fails, point cloud layering may occur, causing the point cloud formed by the laser to be inconsistent with the actual scene, making the use of lidar less reliable and affecting the business related to lidar. to safety hazards.

Contents of the invention

Embodiments of the present disclosure provide a method, device, equipment and storage medium for determining the layering of lidar point clouds.

A first aspect of an embodiment of the present disclosure provides a method for determining the layering of a lidar point cloud. The lidar includes a first field of view and a second field of view. The distance between the first field of view and the second field of view is There are overlapping areas; the method includes:

Obtaining a first candidate point of the LiDAR point cloud, wherein the first candidate point is a point included in the first field of view and located in the overlapping area;

Obtain a second candidate point of the LiDAR point cloud, wherein the second candidate point is a point included in the second field of view and located in the overlapping area;

It is determined whether the point cloud is stratified according to a comparison result between the predetermined characteristics of the first candidate point and the predetermined characteristics of the second candidate point.

In some embodiments according to the first aspect, the method further includes:

In response to determining that the point cloud is stratified, it is determined that at least one of the following anomalies occurs in the lidar:

Displacement of the laser, displacement of the photodetector, abnormal behavior of the MEMS galvanometer and abnormal internal clock.

In some embodiments according to the first aspect, the predetermined characteristics include one or more of the following:

The distance measurement value of the alternative point;

The strength of alternative points;

The normal vector of the alternative point.

In some embodiments according to the first aspect, when the predetermined feature includes a distance measurement value of an alternative point, the predetermined feature according to the first alternative point and the second alternative point Comparison results between predetermined features to determine whether the point cloud is stratified, including:

In response to a difference between the mean of the ranging values of the first candidate point and the mean of the ranging values of the second candidate point being within a first threshold range, determining the point cloud No delamination occurs;

and / or,

In response to a difference between the mean of the ranging values of the first candidate point and the mean of the ranging values of the second candidate point being outside a first threshold range, determining the point cloud Delamination occurs.

In some embodiments according to the first aspect, the difference between the mean of the ranging values in response to the first alternative point and the mean of the ranging values of the second alternative point If the value is within the first threshold range, it is determined that the point cloud is not stratified, including:

In response to a difference between the mean of the ranging values of the first candidate point and the mean of the ranging values of the second candidate point being within a first threshold range, and the first If the difference between the standard deviation of the distance measurement value of the candidate point and the standard deviation of the distance measurement value of the second candidate point is within the second threshold range, it is determined that the point cloud is not stratified. .

In some embodiments according to the first aspect, when the predetermined characteristic includes the intensity of an alternative point, the predetermined characteristic according to the first alternative point and the predetermined characteristic of the second alternative point The comparison results between them determine whether the point cloud is delaminated, including:

In response to the mean of the differences between the intensities of all the first candidate points and the intensities of the nearest neighbor points being within a third threshold range, it is determined that the point cloud is not stratified, wherein, The nearest neighbor point is the point determined from the second alternative points in the second field of view that is closest to each of the first alternative points in the first field of view;

and / or,

In response to a mean value of the differences between the intensities of all the first candidate points and the intensities of the nearest neighbor points being outside a third threshold range, it is determined that the point cloud is stratified, wherein the nearest neighbor point The neighbor point is the point determined from the second candidate points in the second field of view that is closest to each of the first candidate points in the first field of view.

In some embodiments according to the first aspect, when the predetermined feature includes a normal vector of an alternative point, the predetermined feature according to the first alternative point and the predetermined feature of the second alternative point The comparison results between features determine whether the point cloud is stratified, including:

In response to the angle between the first normal vector determined based on the first candidate point and the second normal vector determined based on the second candidate point being within a fourth threshold range, it is determined that the point cloud has not occurred layered;

and / or,

In response to the angle between the first normal vector determined based on the first candidate point and the second normal vector determined based on the second candidate point being outside the fourth threshold range, it is determined that the point cloud is divided layer.

In some embodiments according to the first aspect, determining whether delamination occurs in the point cloud includes:

In response to the comparison result of the N frame point clouds in the M frame point cloud being a predetermined comparison result, it is determined that the point cloud is stratified, wherein the M and N are integers greater than 0, M≥N.

In some embodiments according to the first aspect, the method further includes:

Generate a reference point cloud according to the calibration angle and fixed distance of the lidar;

Determine a first nearest neighbor point in the first field of view in the reference point cloud; select a point whose distance between a point in the second field of view and the first nearest neighbor point is within a predetermined range Determined as a second reference point, wherein the first nearest neighbor point is the point closest to each point in the second field of view determined from the points in the first field of view; Determine a second nearest neighbor point in the second field of view in the reference point cloud; select a point whose distance between the point in the first field of view and the second nearest neighbor point is within a predetermined range. Determined as the first reference point, wherein the second nearest neighbor point is the point closest to each point in the first field of view determined from the points in the second field of view;

Alternatively, the first reference point and the second reference point are determined according to the boundary fitting function of the overlapping area in the reference point cloud.

In some embodiments according to the first aspect, the method further includes:

Determine the serial numbers of the first reference point and the second reference point;

The first candidate point and the second candidate point are determined according to the serial numbers of the first reference point and the second reference point.

In some embodiments according to the first aspect, the method further includes: obtaining a region of interest ROI, in which the first reference point and the second reference point are determined.

A second aspect of the embodiment of the present disclosure provides a device for determining the layering of a lidar point cloud. The lidar includes a first field of view and a second field of view. There is a gap between the first field of view and the second field of view. There are areas of overlap; the means include:

An acquisition module, configured to: acquire a first candidate point of the point cloud of the lidar, where the first candidate point is a point included in the first field of view and located in the overlapping area; Obtain a second candidate point of the LiDAR point cloud, wherein the second candidate point is a point included in the second field of view and located in the overlapping area;

A determination module configured to determine whether the point cloud is stratified according to a comparison result between the predetermined characteristics of the first candidate point and the predetermined characteristics of the second candidate point.

The third aspect of the embodiment of the present disclosure provides a device for determining the layering of lidar point clouds, including:

Memory, which stores computer-executable instructions;

A processor, connected to the memory, is configured to implement the method for determining point cloud layering provided by any solution of the first aspect by executing the computer-executable instructions.

A fourth aspect of the embodiment of the present disclosure provides a computer storage medium, which stores computer-executable instructions; after the computer-executable instructions are executed by a processor, the determination provided by any solution of the first aspect can be realized Point cloud layering method.

The technical solution provided by the embodiments of the present disclosure has a beneficial effect compared with the existing technology: it can be based on the predetermined characteristics of the first candidate point included in the first field of view and located in the overlapping area, and the predetermined characteristics of the first candidate point included in the second field of view. predetermined features of the second candidate point located in the overlapping area, obtain a comparison result between the predetermined features, and determine whether the point cloud is stratified based on the comparison result. Compared with the method of using the human eye to determine the point cloud layering, the accuracy is higher, and after determining that the point cloud is layered, lidar anomalies can be dealt with in a timely manner, which can improve the reliability of lidar work and reduce problems with Safety hazards in lidar-related businesses.

Description of drawings

Figure 1 is a schematic diagram illustrating the cause of a delamination phenomenon provided by an embodiment of the present disclosure;

Figure 2 is a schematic diagram of a normal point cloud provided by an embodiment of the present disclosure;

Figure 3 is a schematic diagram of a layered point cloud provided by an embodiment of the present disclosure;

Figure 4 is a schematic flowchart of a method for determining lidar point cloud layering provided by an embodiment of the present disclosure;

Figure 5 is a schematic diagram of an overlapping area provided by an embodiment of the present disclosure;

Figure 6 is a schematic flowchart of a method for determining lidar point cloud layering provided by an embodiment of the present disclosure;

Figure 7 is a schematic flowchart of a method for determining lidar point cloud layering provided by an embodiment of the present disclosure;

Figure 8 is a schematic flowchart of a method for determining lidar point cloud layering provided by an embodiment of the present disclosure;

Figure 9 is a schematic flow chart of a method for determining lidar point cloud layering provided by an embodiment of the present disclosure;

Figure 10 is a schematic flowchart of a method for determining lidar point cloud layering provided by an embodiment of the present disclosure;

Figure 11 is a schematic flowchart of a method for determining lidar point cloud layering provided by an embodiment of the present disclosure;

Figure 12 is a schematic diagram of determining an overlapping area provided by an embodiment of the present disclosure;

Figure 13 is a schematic flowchart of a method for determining lidar point cloud layering provided by an embodiment of the present disclosure;

Figure 14 is a schematic flow chart of a method for determining lidar point cloud layering provided by an embodiment of the present disclosure;

Figure 15 is a schematic diagram of determining an overlapping area provided by an embodiment of the present disclosure;

Figure 16 is a schematic structural diagram of a device for determining lidar point cloud layering provided by an embodiment of the present disclosure;

Figure 17 is a configuration block diagram of a device for determining lidar point cloud layering provided by an embodiment of the present disclosure.

Detailed ways

In the following description, specific details such as specific system structures and technologies are provided for the purpose of explanation and not limitation, so as to provide a thorough understanding of the embodiments of the present disclosure. However, it will be apparent to those skilled in the art that the present disclosure may be practiced in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, devices, circuits, and methods are omitted so as not to obscure the present disclosure with unnecessary detail.

In order to better understand the embodiments of the present disclosure, first, the relevant application scenarios are described through exemplary embodiments:

LiDAR can obtain information indicating the location of a point in three-dimensional space (for example, its location in the X, Y, and Z planes). Attribute information may also be obtained, such as color attributes (eg, RGB values), texture attributes, intensity attributes, reflectivity attributes, motion-related attributes, modal attributes, and/or various other attributes. In some cases, additional properties can be assigned to the corresponding point, for example, the timestamp when the point was obtained. The points acquired by LiDAR can constitute a "point cloud", which includes a set of points each having associated spatial information and one or more associated attributes. In some cases, a point cloud can include thousands of points, hundreds of thousands of points, millions of points, or even more points. Additionally, in some cases, point clouds can be generated in software. It should be noted that "point" is a "three-dimensional point".

In one embodiment, the laser of a micro-electro-mechanical system (MEMS) scanning lidar is fixedly connected, and light can only propagate along corresponding angles, so that a single laser in the MEMS scanning lidar often Only has a limited field of view. In order to achieve the application requirements of large laser field of view or even full field of view coverage, MEMS scanning lidar can be equipped with Multiple lasers at different angles are installed, and the lasers at different angles are spliced into small fields of view to expand the field of view of the MEMS scanning lidar into a large field of view. At the same time, in order to prevent blind areas between different small fields of view and affect detection accuracy, there are often certain overlapping areas between small fields of view.

Due to reasons such as avalanche photodiode (APD, Avalanche Photo Diode) displacement, abnormal MEMS behavior, and internal clock abnormalities, the point cloud in the overlapping area will be layered.

The fundamental reason for the point cloud layering phenomenon is that the actual ranging does not match the calibrated launch angle. For example, please refer to Figure 1. Lidar O emits a laser OP. Under normal circumstances, the ranging value should be d. However, due to the above reasons for the stratification of the point cloud, the ranging point has changed, from P to P', and the ranging value has correspondingly changed from d to d'. However, the lidar will still be calculated according to the predetermined parameters, and the wrong three-dimensional point P" is calculated based on the abnormal ranging value and the corresponding emission angle in the predetermined parameters. If there is a point in the adjacent field of view that is normal (For example, point P), and the points in another field of view are all like P", a point cloud layering phenomenon will occur. Please refer to Figure 2, which is a point cloud image without point cloud stratification; please refer to Figure 3, which is a point cloud image with point cloud stratification. In this disclosure, “point cloud image” can also be understood as “point cloud”.

In order to illustrate the technical solution of the present invention, specific examples will be described below.

As shown in Figure 4, an embodiment of the present disclosure provides a method for determining the layering of a lidar point cloud. The lidar includes a first field of view and a second field of view, and there is overlap between the first field of view and the second field of view. area; methods include:

Step 41: Obtain the first candidate point of the LiDAR point cloud, where the first candidate point is a point included in the first field of view and located in the overlapping area;

Step 42: Obtain a second candidate point of the LiDAR point cloud, where the second candidate point is a point included in the second field of view and located in the overlapping area;

Step 43: Determine whether the point cloud is stratified according to the comparison result between the predetermined characteristics of the first candidate point and the predetermined characteristics of the second candidate point.

The method for determining lidar point cloud layering provided by the embodiments of the present disclosure can be applied to LiDAR, and the execution of the above steps can be completed by the processing module of LiDAR. However, the method for determining point cloud layering provided by the embodiments of the present disclosure is not limited to application in LiDAR, and can also be applied to various other types of optoelectronic devices or photoelectric sensors including photodetectors or photoelectric receiving circuits. This is not the case here. Make limitations. It should be noted that the method of determining the lidar point cloud layering can also be performed by a host computer connected to the lidar.

In one embodiment, the lidar point cloud may be determined based on multiple fields of view. The field of view may be composed of three-dimensional points (for example, the first candidate point and the second candidate point are both three-dimensional points, and the three-dimensional points may be quantitatively represented by three-dimensional coordinates or other feature information). In the present disclosure, there are mutually overlapping overlapping areas between adjacent fields of view, and the mutually overlapping overlapping areas can also be understood as mutually overlapping overlapping areas. For example, see Figure 5. The point cloud includes field of view A and field B. The overlapping area between field of view A and field B is area C, and area C is the overlapping area. It should be noted that each field of view can correspond to the scanning angle range of the lidar, and the lidar can be divided into different scanning angle ranges. For example, if a scanning angle range of the lidar is 30 degrees, then the 30-degree angle range The scanning area can correspond to a field of view. Different scanning angle ranges can overlap, and thus different fields of view can also overlap. It should be noted that the scanning angle may include an azimuth angle and an elevation angle. The scanning angle range in the above example may be an azimuth angle range and/or an elevation angle range, which is not limited here.

In one embodiment, each field of view can be represented by a point set composed of three-dimensional points contained in the field of view. For example, if the first field of view includes multiple three-dimensional points, then the multiple three-dimensional points in the first field of view can constitute a point set, such as the P1 point set, that is, the first field of view can be represented by this point set. The three-dimensional point associated with the first field of view in the present disclosure may be any three-dimensional point in the P1 point set. For another example, if the second field of view includes multiple three-dimensional points, the multiple three-dimensional points in the second field of view can constitute a point set, such as the P2 point set, that is, the second field of view can be represented by this point set. The three-dimensional point associated with the first field of view in the present disclosure may be any three-dimensional point in the P2 point set. Here, the point set corresponding to the visual field can be used for the operation implementation of the disclosed solution. In one embodiment, the first field of view and the second field of view may be adjacent fields of view. It should be noted that the first field of view and the second field of view do not specifically refer to any two fields of view. It can be understood that the first field of view and the second field of view can be any adjacent fields of view in the field of view, and there is no limitation here.

In one embodiment, the predetermined characteristics include one or more of the following:

The distance measurement value of the alternative point;

The strength of alternative points;

The normal vector of the alternative point.

In one embodiment, the first candidate point and the second candidate point are obtained; according to a predetermined period, according to the predetermined characteristics of the first candidate point and The comparison results between the predetermined characteristics of the second candidate point periodically determine whether the point cloud is stratified. The predetermined period may be determined based on the required abnormal response delay. For example, in response to the required abnormal response delay being less than the delay threshold, it is determined that the predetermined period is less than the period threshold; or in response to the required abnormal response delay being greater than the delay threshold, it is determined that the predetermined period is greater than the period threshold. In this way, the predetermined period can be adapted to the required exception response delay.

In one embodiment, a first candidate point is obtained, wherein the first candidate point is a point included in the first field of view and located in an overlapping area; a second candidate point is obtained, wherein the second candidate point is a point included in the second field of view and located in the overlapping area; based on the comparison result between the predetermined characteristics of the first candidate point and the predetermined characteristics of the second candidate point, it is determined whether the point cloud is stratified. In response to determining that the point cloud is stratified, it is determined that the lidar has at least one of the following abnormalities: displacement of the laser, displacement of the photodetector, abnormal MEMS behavior, and abnormal internal clock. In this way, after it is determined that the point cloud is stratified, abnormal processing can be carried out in time to improve the reliability of lidar work. For example, in response to determining that the point cloud is stratified, prompt information indicating that the above-mentioned abnormality occurs in the lidar is output. It should be noted that the photodetector can be an APD or a single photon avalanche diode (SPAD, Single Photon Avalanche Diode), etc.

In one embodiment, a first candidate point is obtained, wherein the first candidate point is a point included in the first field of view and located in an overlapping area; a second candidate point is obtained, wherein the second candidate point is a point included in the second field of view and located in the overlapping area; based on the comparison result between a single predetermined feature of the first candidate point and a single predetermined feature of the second candidate point, it is determined whether the point cloud is stratified.

In one embodiment, a first candidate point is obtained, wherein the first candidate point is a point included in the first field of view and located in an overlapping area; a second candidate point is obtained, wherein the second candidate point is a point included in the second field of view and located in the overlapping area; based on the comparison result between the ranging value of the first candidate point and the ranging value of the second candidate point, it is determined whether the point cloud is stratified.

In one embodiment, a first candidate point is obtained, wherein the first candidate point is a point included in the first field of view and located in an overlapping area; a second candidate point is obtained, wherein the second candidate point is a point included in the second field of view and located in the overlapping area; based on the difference between the mean value of the distance measurement value of the first candidate point and the mean value of the range measurement value of the second candidate point, determine whether the point cloud Delamination occurs. For example, in response to the difference being within the first threshold range, it is determined that the point cloud is not stratified; or in response to the difference being outside the first threshold range, it is determined that the point cloud is stratified.

In one embodiment, a first candidate point is obtained, wherein the first candidate point is a point included in the first field of view and located in an overlapping area; a second candidate point is obtained, wherein the second candidate point is a point included in the second field of view and located in the overlapping area; the difference between the mean value of the ranging values according to the first candidate point and the mean value of the ranging values of the second candidate point, and according to the first The difference between the standard deviation of the ranging value of the candidate point and the standard deviation of the ranging value of the second candidate point determines whether the point cloud is stratified. Exemplarily, in response to the two differences being within a predetermined range (for example, the difference between the mean is within the first threshold range, and the difference between the standard deviation is within the second threshold range), it is determined that the point cloud is not stratified. ; Or, in response to both the difference values being outside the predetermined range, it is determined that the point cloud is stratified, or in response to one of the two difference values being outside the predetermined range, it is determined that the point cloud is stratified.

In one embodiment, a first candidate point is obtained, wherein the first candidate point is a point included in the first field of view and located in an overlapping area; a second candidate point is obtained, wherein the second candidate point is a point included in the second field of view and located in the overlapping area; in response to the mean value of the difference between the intensity of all first candidate points and the intensity of the nearest neighbor point being within the third threshold range, it is determined that the point cloud has not Stratification occurs where the nearest neighbor point is the point determined from the second candidate points in the second field of view that is closest to each first candidate point in the first field of view.

In one embodiment, a first candidate point is obtained, wherein the first candidate point is a point included in the first field of view and located in an overlapping area; a second candidate point is obtained, wherein the second candidate point is a point included in the second field of view and located in the overlapping area; in response to the mean value of the difference between the intensity of all first candidate points and the intensity of the nearest neighbor point being outside the third threshold range, it is determined that the point cloud occurs Hierarchical, wherein the nearest neighbor point is the point determined from the second candidate points in the second field of view that is closest to each first candidate point in the first field of view.

In one embodiment, a first candidate point is obtained, wherein the first candidate point is a point included in the first field of view and located in an overlapping area; a second candidate point is obtained, wherein the second candidate point is a point included in the second field of view and located in the overlapping area; based on the angle between the first normal vector of the first candidate point and the second normal vector of the second candidate point, determine whether the point cloud is divided. layer. Exemplarily, in response to the angle between the first normal vector determined based on the first candidate point and the second normal vector determined based on the second candidate point being within a fourth threshold range, it is determined that the point cloud is not delaminated. .

In one embodiment, a first candidate point is obtained, wherein the first candidate point is a point included in the first field of view and located in an overlapping area; a second candidate point is obtained, wherein the second candidate point is a point included in the second field of view and located in the overlapping area; based on the angle between the first normal vector of the first candidate point and the second normal vector of the second candidate point, determine whether the point cloud is divided. layer. Exemplarily, in response to the angle between the first normal vector determined based on the first candidate point and the second normal vector determined based on the second candidate point being outside a fourth threshold range, it is determined that the point cloud is stratified.

In one embodiment, a first candidate point is obtained, wherein the first candidate point is a point included in the first field of view and located in an overlapping area; a second candidate point is obtained, wherein the second candidate point is a point included in the second field of view and located in the overlapping area; based on the comparison results between the multiple predetermined features of the first candidate point and the multiple predetermined features of the second candidate point, it is determined whether the point cloud is divided. layer. Among them, the plurality of predetermined features include the following: the distance measurement value of the candidate point; the intensity of the candidate point; and the normal vector of the candidate point. Here, determining whether the point cloud is delaminated based on the comparison results between multiple predetermined features can improve the accuracy of determining whether the point cloud is delaminated.

The technical solution provided by the embodiments of the present disclosure has a beneficial effect compared with the existing technology: it can be based on the predetermined characteristics of the first candidate point included in the first field of view and located in the overlapping area, and the predetermined characteristics of the first candidate point included in the second field of view. predetermined features of the second candidate point located in the overlapping area, obtain a comparison result between the predetermined features, and determine whether the point cloud is stratified based on the comparison result. Compared with the method of using the human eye to determine the point cloud layering, the accuracy is higher, and after determining that the point cloud is layered, lidar anomalies can be dealt with in a timely manner, which can improve the reliability of lidar work and reduce problems with Safety hazards in lidar-related businesses.

It should be noted that those skilled in the art can understand that the methods provided in the embodiments of the present disclosure can be executed alone or together with some methods in the embodiments of the present disclosure or some methods in related technologies.

As shown in Figure 6, an embodiment of the present disclosure provides a method for determining the layering of a lidar point cloud. The lidar includes a first field of view and a second field of view, and there is overlap between the first field of view and the second field of view. area; methods include:

Step 61: In response to determining that the point cloud is stratified, it is determined that the lidar has at least one of the following abnormalities: displacement of the laser, displacement of the photodetector, abnormal MEMS behavior, and abnormal internal clock.

In one embodiment, the first candidate point of the point cloud of the lidar is obtained, where the first candidate point is a point included in the first field of view and located in the overlapping area; and the first candidate point of the point cloud of the lidar is obtained. Two candidate points, wherein the second candidate point is a point included in the second field of view and located in the overlapping area; based on the comparison between the predetermined characteristics of the first candidate point and the predetermined characteristics of the second candidate point As a result, it is determined whether the point cloud is delaminated. In response to determining that the m frame point clouds in the n frame point clouds are stratified, it is determined that the lidar has at least one of the following abnormalities: displacement of the laser, displacement of the photodetector, abnormal MEMS behavior, and abnormal internal clock. In this way, after it is determined that the point cloud is stratified, abnormal processing can be carried out in time to improve the reliability of lidar work. Here, n and m are positive integers, and the ratio of m and n is greater than a predetermined threshold. In one embodiment, in response to the required accuracy of abnormal handling being greater than the accuracy threshold, it is determined that the predetermined threshold is greater than the reference value; or in response to the required accuracy of abnormal handling being less than the accuracy threshold, it is determined that the predetermined threshold is less than the reference value.

It should be noted that those skilled in the art can understand that the methods provided in the embodiments of the present disclosure can be executed alone or together with some methods in the embodiments of the present disclosure or some methods in related technologies.

As shown in Figure 7, the embodiment of the present disclosure provides a method for determining the layering of lidar point clouds. The lidar includes a first field of view and a second field of view, and there is overlap between the first field of view and the second field of view. Area; when the predetermined features include ranging values of candidate points, the method includes:

Step 71: In response to the difference between the mean value of the ranging value of the first candidate point and the mean value of the ranging value of the second candidate point being within the first threshold range, it is determined that the point cloud is not stratified;

and / or,

In response to the difference between the mean value of the ranging values of the first candidate point and the mean value of the ranging values of the second candidate point being outside the first threshold range, it is determined that the point cloud is stratified.

In one embodiment, in response to a difference between the mean value of the ranging values of the first candidate point and the mean value of the ranging values of the second candidate point being within a first threshold range, and the value of the mean value of the ranging value of the first candidate point If the difference between the standard deviation of the ranging value and the standard deviation of the ranging value of the second candidate point is within the second threshold range, it is determined that the point cloud is not stratified.

For specific description of step 71 in the embodiment of the present disclosure, please refer to the description of step 41, step 42 and step 43, which will not be described again here.

It should be noted that those skilled in the art can understand that the methods provided in the embodiments of the present disclosure can be executed alone or together with some methods in the embodiments of the present disclosure or some methods in related technologies.

As shown in Figure 8, an embodiment of the present disclosure provides a method for determining the layering of a lidar point cloud. The lidar includes a first field of view and a second field of view, and there is overlap between the first field of view and the second field of view. Area; when the predetermined characteristics include the intensity of candidate points, the method includes:

Step 81: In response to the mean value of the difference between the intensity of all first candidate points and the intensity of the nearest neighbor point being within the third threshold range, determine that the point cloud has not been stratified, where the nearest neighbor point is from the second The point closest to each first candidate point in the first field of view determined among the second candidate points in the field of view;

and / or,

Determining the point cloud in response to a mean of differences between the intensities of all first candidate points and the intensities of nearest neighbor points being outside a third threshold range Stratification occurs, wherein the nearest neighbor point is the point determined from the second candidate points in the second field of view that is closest to each first candidate point in the first field of view;

and / or,

In response to the mean of the differences between the intensities of all second candidate points and the intensities of the nearest neighbor points being within the third threshold range, it is determined that the point cloud is not stratified, wherein the nearest neighbor points are from the first field of view. The point closest to the determined distance between the first candidate point and each second candidate point in the second field of view;

and / or,

In response to the mean value of the difference between the intensities of all second candidate points and the intensity of the nearest neighbor point being outside the third threshold range, it is determined that the point cloud is stratified, wherein the nearest neighbor point is the third point from the first field of view. The point closest to each second candidate point in the second field of view is determined among the candidate points.

For specific description of step 81 in the embodiment of the present disclosure, please refer to the description of step 41, step 42 and step 43, which will not be described again here.

It should be noted that those skilled in the art can understand that the methods provided in the embodiments of the present disclosure can be executed alone or together with some methods in the embodiments of the present disclosure or some methods in related technologies.

As shown in Figure 9, the embodiment of the present disclosure provides a method for determining the layering of lidar point clouds. The lidar includes a first field of view and a second field of view, and there is overlap between the first field of view and the second field of view. Area; when the predetermined feature includes the normal vector of the candidate point, the method includes:

Step 91: In response to the angle between the first normal vector determined based on the first candidate point and the second normal vector determined based on the second candidate point being within the fourth threshold range, it is determined that the point cloud is not stratified;

and / or,

In response to the angle between the first normal vector determined based on the first candidate point and the second normal vector determined based on the second candidate point being outside the fourth threshold range, it is determined that the point cloud is stratified.

For the specific description of step 91 in the embodiment of the present disclosure, please refer to the description of step 41, step 42, and step 43, which will not be described again here.

It should be noted that those skilled in the art can understand that the methods provided by the embodiments of the present disclosure can be executed alone or together with some methods in the embodiments of the present disclosure or some methods in related technologies.

As shown in Figure 10, an embodiment of the present disclosure provides a method for determining the layering of a lidar point cloud. The lidar includes a first field of view and a second field of view, and there is overlap between the first field of view and the second field of view. area; methods include:

Step 101: In response to the comparison result of the N frame point clouds in the M frame point cloud being a predetermined comparison result, it is determined that the point cloud is stratified, where M and N are integers greater than 0, and M≥N.

In one embodiment, the first candidate point of the point cloud of the lidar is obtained, where the first candidate point is a point included in the first field of view and located in the overlapping area; and the first candidate point of the point cloud of the lidar is obtained. Two alternative points, wherein the second alternative point is a point included in the second field of view and located in the overlapping area; based on the comparison between the predetermined characteristics of the first alternative point and the predetermined characteristics of the second alternative point As a result, it is determined whether the point cloud is delaminated. For example, in response to the comparison result of the N frame point clouds in the M frame point cloud being a predetermined comparison result, it is determined that the point cloud is stratified, where M and N are integers greater than 0, and M≥N.

In one embodiment, in response to the difference between the average of the ranging values of the first candidate points and the average of the ranging values of the second candidate points in the N frame point clouds in the M frame point clouds, the first Within the threshold range, it is determined that the comparison result is not the predetermined comparison result.

In one embodiment, in response to the difference between the average of the ranging values of the first candidate points and the average of the ranging values of the second candidate points in the N frame point clouds in the M frame point clouds, the first If the value is outside the threshold range, the comparison result is determined to be the predetermined comparison result.

In one embodiment, in response to a mean value of the difference between the intensity of each first candidate point and the intensity of the nearest neighbor point of the N frame point clouds in the M frame point cloud being within a third threshold range, determining the comparison The result is not a predetermined comparison result, wherein the nearest neighbor point is the point determined from the second candidate points in the second field of view that is closest to each first candidate point in the first field of view.

In one embodiment, in response to a mean value of the difference between the intensity of each first candidate point and the intensity of the nearest neighbor point of the N frame point clouds in the M frame point cloud being outside a third threshold range, determining the comparison The result is a predetermined comparison result, wherein the nearest neighbor point is the point determined from the second candidate points in the second field of view that is closest to each first candidate point in the first field of view.

In one embodiment, in response to a mean value of the difference between the intensity of each second candidate point and the intensity of the nearest neighbor point of the N frame point clouds in the M frame point cloud being within a third threshold range, determining the comparison The result is not a predetermined comparison result, wherein the nearest neighbor point is the point determined from the first candidate points in the first field of view that is closest to each second candidate point in the second field of view.

In one embodiment, in response to a mean value of the difference between the intensity of each second candidate point and the intensity of the nearest neighbor point of the N frame point clouds in the M frame point cloud being outside a third threshold range, determining the comparison The result is a predetermined comparison result, in which the nearest neighbor point is the point determined from the first candidate points in the first field of view that is closest to each second candidate point in the second field of view.

In one embodiment, in response to the N frame point cloud in the M frame point cloud, the response is between a first normal vector determined based on the first candidate point and a second normal vector determined based on the second candidate point. The angle is within the fourth threshold range, and it is determined that the comparison result is not the predetermined comparison result.

In one embodiment, in response to the N frame point cloud in the M frame point cloud, the response is between a first normal vector determined based on the first candidate point and a second normal vector determined based on the second candidate point. If the angle is outside the threshold range, it is determined that the point cloud is stratified, and the comparison result is determined to be the predetermined comparison result.

It should be noted that those skilled in the art can understand that the methods provided in the embodiments of the present disclosure can be executed alone or together with some methods in the embodiments of the present disclosure or some methods in related technologies.

As shown in Figure 11, the embodiment of the present disclosure provides a method for determining the layering of lidar point clouds. The lidar includes a first field of view and a second field of view, and there is overlap between the first field of view and the second field of view. area; methods include:

Step 111. Generate a reference point cloud based on the calibration angle and fixed distance of the lidar;

Step 112: Determine the first nearest neighbor point in the first field of view in the reference point cloud; determine a point whose distance between the point in the second field of view and the first nearest neighbor point is within a predetermined range as the second reference points, wherein the first nearest neighbor point is the point closest to each point in the second field of view determined from the points in the first field of view;

Determine the second nearest neighbor point in the second field of view in the reference point cloud; determine a point whose distance between the point in the first field of view and the second nearest neighbor point is within a predetermined range as the first reference point, where , the second nearest neighbor point is the point closest to each point in the first field of view determined from the points in the second field of view;

Alternatively, the first reference point and the second reference point are determined according to the boundary fitting function of the overlapping area in the reference point cloud.

In some embodiments, the calibration angle is the calibration information indicated in the calibration file of the lidar, and the fixed distance is the distance set according to the application scenario.

It should be noted that the calibration information may be information about the parameters of the lidar under actual working conditions when the lidar leaves the factory; the calibration information may at least include scanning angle information. A reference point cloud can be generated based on the scan angle information and a predetermined distance. The reference point cloud may also be called an auxiliary point cloud.

In one embodiment, a reference point cloud is generated based on the calibration angle and fixed distance of the lidar; the first nearest neighbor point in the first field of view is determined in the reference point cloud; and the point in the second field of view is compared with the first nearest point cloud. A point whose distance between adjacent points is within a predetermined range is determined as the second reference point, wherein the first nearest neighbor point is determined from the points in the first field of view and each point in the second field of view. the closest point to A first reference point, wherein the second nearest neighbor point is the point closest to each point in the first field of view determined from the points in the second field of view; based on the first reference point and the second The serial number of the reference point determines the first candidate point and the second candidate point. It is determined whether the point cloud is stratified according to a comparison result between the predetermined characteristics of the first candidate point and the predetermined characteristics of the second candidate point.

It should be noted that points in the field of view can all correspond to unique serial numbers to achieve fast traversal operations. In the present disclosure, after the first reference point and the second reference point are determined, the serial numbers of the first reference point and the second reference point can be determined. In this way, after the serial numbers are determined, the first candidate point can be quickly determined based on the serial numbers. and a second alternative point.

In one embodiment, a single frame reference point cloud can be generated based on the calibration information (including scanning angle information) and a predetermined distance; traverse the three-dimensional points in the field of view A, and use KD-Tree to search in the adjacent field of view B with A A nearest neighbor point of each three-dimensional point in the field of view; if the Euclidean distance between the three-dimensional point in the field of view A and the nearest neighbor point is less than a predetermined value, the three-dimensional point in the field of view A can be determined The overlapping areas located in different fields of view are the reference points used to determine candidate points. For example, please refer to Figure 12. Taking the three-dimensional point P1 in the upper left corner of the field of view A and the three-dimensional point P2 in the upper right corner as an example, calculate the Euclidean distances from P1 and P2 to the nearest neighbor point in the field of view B. If P1 is in B The distance of the nearest neighbor point is much greater than the set threshold X (for example, 0.02 meters). Then P1 does not belong to the overlapping area of field of view A and field of view B. If the distance from point P2 to the nearest neighbor point in field of view B is within the threshold range, then P2 belongs to the overlapping area of field of view A and field of view B, as The reference point used to determine alternative points. It should be noted that in the above example, if A is the first field of view, B is the second field of view, the corresponding nearest neighbor point is the second nearest neighbor point, and the corresponding reference point is the first reference point; if A is The second field of view, then B is the first field of view, the corresponding nearest neighbor point is the first nearest neighbor point, and the corresponding reference point is the second reference point.

In one embodiment, a reference point cloud is generated based on the calibration angle and fixed distance of the lidar; the first reference point and the second reference point are determined from the reference point cloud based on the boundary fitting function of the overlapping area. The first reference point is determined based on the serial numbers of the first reference point and the second reference point. Alternative point and second alternative point. It is determined whether the point cloud is stratified according to a comparison result between the predetermined characteristics of the first candidate point and the predetermined characteristics of the second candidate point.

In one embodiment, a single-frame reference point cloud (or auxiliary point cloud) may be generated based on the calibration information, where the reference point cloud contains multiple fields of view; since the edge point number of each field of view (which may be the boundary of the field of view) The point number) has been determined when the lidar is calibrated at the factory. Therefore, the edge points of each field of view can be determined based on the edge point numbers determined based on the calibration information. A set of different fields of view can be fitted based on the information of the edge points. The system of equations of the boundary curve, so that by substituting the coordinates of each point in the field of view into the system of equations, it can be determined whether the point belongs to the overlapping area of different fields of view. If the point belongs to the overlapping area, it is the reference point. In this way, the serial number of the reference point in the overlapping area can be determined. This serial number can be used to quickly determine the first alternative point and the second alternative point. It should be noted that if the boundary fitting function corresponds to the first field of view, the reference point is the first reference point; or, if the boundary fitting function corresponds to the second field of view, the reference point is the second reference point.

In one embodiment, the upper, lower, left and right boundary curves of each field of view can be fitted using a cubic function. For example, the cubic function may be: Ax3+Bx2+Cx+D=0. In this way, the first reference point or the second reference point can be determined through a cubic function equation set determined by multiple cubic functions in different fields of view.

It should be noted that those skilled in the art can understand that the methods provided in the embodiments of the present disclosure can be executed alone or together with some methods in the embodiments of the present disclosure or some methods in related technologies.

As shown in Figure 13, the embodiment of the present disclosure provides a method for determining the layering of lidar point clouds. The lidar includes a first field of view and a second field of view, and there is overlap between the first field of view and the second field of view. area; methods include:

Step 131: Obtain the region of interest ROI, and determine the first reference point and the second reference point in the ROI.

In one embodiment, the region of interest (ROI, Region of Interest) may be determined based on the emission angle of the lidar, and the first reference point and the second reference point are determined only in the ROI. For example, the ROI is the area where the horizontal emission angle is between the minimum value of the Azimuth azimuth angle of the B field of view and the maximum value of the Azimuth azimuth angle of the A field of view, then the overlap of the A field of view and the B field of view can be determined based on this ROI The horizontal emission angle of the three-dimensional point in the area is between the minimum value of the Azimuth azimuth angle of the B field of view and the maximum value of the Azimuth azimuth angle of the A field of view. Based on the method of step 112, it is only necessary to traverse the point cloud in the ROI to determine the first reference point and the second reference point. In this way, the amount of calculation can be reduced, and therefore the determination of the first reference point and the second reference point can be accelerated.

In one embodiment, a region of interest ROI is obtained, and a first reference point and a second reference point are determined in the ROI. The first candidate point and the second candidate point are determined based on the serial numbers of the first reference point and the second reference point. It is determined whether the point cloud is stratified according to a comparison result between the predetermined characteristics of the first candidate point and the predetermined characteristics of the second candidate point.

In order to better understand the embodiments of the present disclosure, the following further describes the embodiments of the present disclosure through an exemplary embodiment:

See Figure 14. This example provides a method to determine the layering of lidar point clouds, including:

Step a1: Based on the laser lidar factory calibration file (for example, angle calibration file) and the fixed distance, determine all three-dimensional points belonging to the point sets of the two fields of view in each overlapping area of the two fields of view. Among them, each three-dimensional point corresponds to a serial number. For example, you can generate a frame of reference point cloud based on the angle calibration file and a fixed distance, traverse the point cloud of a certain field of view, and use KD-Tree to search for the nearest neighbor point of each three-dimensional point in the adjacent field of view. If the Euclidean distance between the three-dimensional point and the nearest neighbor point in the adjacent field of view is less than the set threshold, the point is considered to belong to the overlapping field of view area. Please refer to Figure 12 again. Taking the three-dimensional point P1 in the upper left corner of the field of view A and the three-dimensional point P2 in the upper right corner as an example, calculate the Euclidean distances from P1 and P2 to the nearest neighbor point in the field of view B. Obviously, the distance from P1 to the nearest neighbor in B is The distance is much greater than the threshold 0.02 meters we set, so P1 does not belong to the AB field of view overlap area point, and the distance from point P2 to the nearest neighbor point in B is within the threshold range, so P2 belongs to the AB field of view overlap area point set , the points included in this point set are reference points (for example, the first reference point or the second reference point), and each reference point corresponds to a sequence number. It should be noted that step a1 only needs to be executed once when the laser lidar is powered on or the detection algorithm is started. In order to speed up the search, an ROI can first be determined based on the emission angle. For example, the point horizontal emission angle of the overlap area of the A field of view and the B field of view is between the minimum Azimuth azimuth angle of the B field of view and the maximum Azimuth azimuth angle of the A field of view. Then, determine the reference point only in the ROI.

Step a2: During the normal use of the laser lidar, calculate the eigenvalues of the point sets P1 and P2 belonging to the two fields of view in each overlapping area. Taking Figure 15 as an example, there are a total of 10 main overlapping areas, as shown in Figure 15. Take the field of view overlapping area numbered ① as an example. The upper field of view point set in this overlapping area is P1, and the lower field of view point set is P2. The statistics of the two point sets P1 belonging to the two fields of view in this overlapping area are And the mean and/or standard deviation of the distance values in P2 (here, after sorting, delete the largest and smallest 10% of the distance values) as the feature value A. It should be noted that the eigenvalues that can also be used include intensity (eigenvalue B) and the normal vector of the point cloud (eigenvalue C).

Step a3: Determine whether the single frame point cloud is delaminated based on the feature value. It can be judged based on a single eigenvalue. Taking eigenvalue A as an example, if the difference between the mean distance value and the standard deviation of the two sets of point sets P1 and P2 in an overlapping area are both within the set threshold, it is considered There is no stratification phenomenon in this overlapping area, otherwise it is judged that stratification phenomenon has occurred in this overlapping area of this frame (the degree of dispersion is similar, but the mean value is quite different).

The feature value can also be the value of feature B. Feature B is the intensity information of the point cloud. It belongs to the original measurement information like the distance. In theory, the intensity of similar points in the overlapping area should also be consistent. The judgment method is: traverse each point in P1, find the corresponding nearest neighbor point in P2, record the intensity difference between the two points, and find the mean of all differences after the traversal. If it exceeds the threshold, it is judged that the point cloud is stratified.

The feature value can also be the value of feature C, which is the normal vector of the point cloud. Set a threshold based on the angle between the normal vectors to determine whether delamination occurs, and calculate the angle between the normal vectors x1 and x2 of the two sets of point clouds P1 and P2. Angle θ, if θ exceeds the threshold, the point cloud is judged to be stratified.

The above three feature values can also be combined for judgment. The solution of combining multiple features can determine whether the point cloud is stratified according to whether two of the three features meet the conditions for stratification, thus improving the accuracy of the judgment.

Step a4: Count the proportion of frames with layered points in multiple frames. If the number of layered point clouds exceeds a certain threshold, an exception will be reported. For example, if at least 40 frames of 50 consecutive point cloud frames are delaminated, it will be judged that the device is delaminated and an exception will be reported.

As shown in Figure 16, an embodiment of the present disclosure provides a device for determining the layering of a lidar point cloud. The lidar includes a first field of view and a second field of view. There is overlap between the first field of view and the second field of view. Area; devices include:

The acquisition module 161 is used to: acquire the first candidate point of the laser radar point cloud, where the first candidate point is a point included in the first field of view and located in the overlapping area; obtain the first candidate point of the laser radar point cloud. A second candidate point, wherein the second candidate point is a point included in the second field of view and located in the overlapping area;

The determination module 162 is configured to determine whether the point cloud is stratified according to the comparison result between the predetermined characteristics of the first candidate point and the predetermined characteristics of the second candidate point.

As shown in Figure 17, the embodiment of the present disclosure provides a device 1700 for determining LiDAR point cloud layering, including:

Memory 1704, which stores computer-executable instructions;

The processor 1703 is connected to the memory 1704 and is used to implement the failure detection method of the signal receiving component provided by any of the foregoing technical solutions by executing computer executable instructions. For example, the processor 1703 can implement this method by executing the executable instructions. Expose any method.

Device 1700 may be any type of general or special purpose computing device, such as a desktop computer, laptop computer, server, mainframe computer, cloud-based computer, tablet computer, wearable device, vehicle electronics, etc. As shown in FIG. 17 , device 1700 includes an input/output (I/O) interface 1701 , a network interface 1702 , a memory 1704 and a processor 1703 .

I/O interface 1701 is a collection of components that can receive input from and/or provide output to the user. I/O interface 1701 may include, but is not limited to, buttons, keyboards, keypads, LCD displays, LED displays, or other similar display devices, including display devices with touch screen capabilities that enable interaction between the user and the device.

Network interface 1702 may include various adapters and circuitry implemented in software and/or hardware to enable communication with the lidar using wired or wireless protocols. The wired protocol is, for example, any one or more of a serial port protocol, a parallel port protocol, an Ethernet protocol, a USB protocol or other wired communication protocols. The wireless protocol is, for example, any IEEE 802.11 Wi-Fi protocol, cellular network communication protocol, etc.

Memory 1704 includes a single memory or one or more memories or storage locations, including but not limited to random access memory (RAM), dynamic random access memory (DRAM), static random access memory (SRAM), read only memory (ROM) ), EPROM, EEPROM, flash memory, logic blocks of FPGA, hard disk, or any other layer of the memory hierarchy. Memory 1704 may be used to store any type of instructions, software, or algorithms, including instructions 1705 for controlling the general functionality and operation of device 1700 .

Processor 1703 controls the general operation of device 1700. The processor 1703 may include, but is not limited to, a CPU, a hardware microprocessor, a hardware processor, a multi-core processor, a single-core processor, a microcontroller, an application specific integrated circuit (ASIC), a DSP, or other similar processing device capable of executing Any type of instructions, algorithms, or software for controlling the operation and functionality of device 1700 of the embodiments described in this disclosure. Processor 1703 may be various implementations of digital circuitry, analog circuitry, or mixed-signal (a combination of analog and digital) circuitry that perform functions in a computing system. Processor 1703 may include, for example, a portion or circuit such as an integrated circuit (IC), a separate processor core, an entire processor core, a separate processor, a programmable hardware device such as a field programmable gate array (FPGA), and/or Systems that include multiple processors.

The processor 1703 and the memory 1704 may be connected through a communication interface such as a bus 1706.

Embodiments of the present disclosure also provide a computer storage medium that stores computer-executable instructions; the computer-executable instructions are stored in the computer storage medium; After the execution instruction is executed by the processor, the method for determining the lidar point cloud layer provided by any of the foregoing technical solutions can be implemented. For example, the processor can implement any method of the present disclosure by executing the executable instruction.

Computer-executable instructions in the computer-readable storage medium or program product according to embodiments of the present disclosure may be configured to perform operations corresponding to the above-described apparatus and method embodiments. When referring to the above-described apparatus and method embodiments, the embodiments of the computer-readable storage medium or program product will be clear to those skilled in the art, and therefore will not be described again. Computer-readable storage media and program products for carrying or including the computer-executable instructions described above are also within the scope of the present disclosure. Such storage media may include, but are not limited to, floppy disks, optical disks, magneto-optical disks, memory cards, memory sticks, and the like.

Those skilled in the art can understand that the size of the sequence numbers of each step in the above embodiments does not mean the order of execution. The execution order of each process should be determined by its functions and internal logic, and should not be determined by the implementation process of the embodiments of the present disclosure. constitute any limitation.

The above-described embodiments are only used to illustrate the technical solutions of the present disclosure, but not to limit them; although the present disclosure has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still implement the foregoing implementations. Modifications are made to the technical solutions described in the examples, or equivalent substitutions are made to some of the technical features; these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of each embodiment of the present disclosure, and should be included in within the scope of this disclosure.

Claims (24)

1. A method for determining the layering of lidar point clouds, wherein the lidar includes a first field of view and a second field of view, and there is an overlapping area between the first field of view and the second field of view; The methods include:

   Obtaining a first candidate point of the LiDAR point cloud, wherein the first candidate point is a point included in the first field of view and located in the overlapping area;

   Obtain a second candidate point of the LiDAR point cloud, wherein the second candidate point is a point included in the second field of view and located in the overlapping area;

   It is determined whether the point cloud is stratified according to a comparison result between the predetermined characteristics of the first candidate point and the predetermined characteristics of the second candidate point.
2. The method of claim 1, further comprising:

   In response to determining that the point cloud is stratified, it is determined that at least one of the following anomalies occurs in the lidar:

   Displacement of the laser, displacement of the photodetector, abnormal behavior of the MEMS galvanometer and abnormal internal clock.
3. The method according to claim 1, wherein the predetermined characteristics include one or more of the following:

   The distance measurement value of the alternative point;

   The strength of alternative points;

   The normal vector of the alternative point.
4. The method of claim 3, wherein when the predetermined feature includes a distance measurement value of an alternative point, the predetermined feature according to the first alternative point and the second alternative point are Comparison results between predetermined features to determine whether the point cloud is stratified, including:

   In response to a difference between the mean of the ranging values of the first candidate point and the mean of the ranging values of the second candidate point being within a first threshold range, determining the point cloud No delamination occurs;

   and / or,

   In response to a difference between the mean of the ranging values of the first candidate point and the mean of the ranging values of the second candidate point being outside a first threshold range, determining the point cloud Delamination occurs.
5. The method of claim 4, wherein the difference between the mean of the ranging values in response to the first alternative point and the mean of the ranging values of the second alternative point If the value is within the first threshold range, it is determined that the point cloud is not stratified, including:

   In response to a difference between the mean of the ranging values of the first candidate point and the mean of the ranging values of the second candidate point being within a first threshold range, and the first If the difference between the standard deviation of the distance measurement value of the candidate point and the standard deviation of the distance measurement value of the second candidate point is within the second threshold range, it is determined that the point cloud is not stratified. .
6. The method of claim 3, wherein when the predetermined feature includes an intensity of a candidate point, the predetermined feature according to the first candidate point and the predetermined feature of the second candidate point The comparison results between them determine whether the point cloud is delaminated, including:

   In response to the mean of the differences between the intensities of all the first candidate points and the intensities of the nearest neighbor points being within a third threshold range, it is determined that the point cloud is not stratified, wherein, The nearest neighbor point is the point determined from the second alternative points in the second field of view that is closest to each of the first alternative points in the first field of view;

   and / or,

   In response to a mean value of the differences between the intensities of all the first candidate points and the intensities of the nearest neighbor points being outside a third threshold range, it is determined that the point cloud is stratified, wherein the nearest neighbor point The neighbor point is the point determined from the second candidate points in the second field of view that is closest to each of the first candidate points in the first field of view.
7. The method of claim 3, wherein when the predetermined feature includes a normal vector of an alternative point, the predetermined feature based on the first alternative point and the predetermined feature of the second alternative point The comparison results between features determine whether the point cloud is stratified, including:

   In response to the angle between the first normal vector determined based on the first candidate point and the second normal vector determined based on the second candidate point being within a fourth threshold range, it is determined that the point cloud has not occurred layered;

   and / or,

   In response to a sandwich between a first normal vector determined based on the first alternative point and a second normal vector determined based on the second alternative point If the angle is outside the fourth threshold range, it is determined that the point cloud is stratified.
8. The method according to any one of claims 1 to 7, wherein determining whether delamination occurs in the point cloud includes:

   In response to the comparison result of the N frame point clouds in the M frame point cloud being a predetermined comparison result, it is determined that the point cloud is stratified, wherein the M and N are integers greater than 0, M≥N.
9. The method according to any one of claims 1 to 7, wherein the method further includes:

   Generate a reference point cloud according to the calibration angle and fixed distance of the lidar;

   Determine a first nearest neighbor point in the first field of view in the reference point cloud; select a point whose distance between a point in the second field of view and the first nearest neighbor point is within a predetermined range Determined as a second reference point, wherein the first nearest neighbor point is the point closest to each point in the second field of view determined from the points in the first field of view; Determine a second nearest neighbor point in the second field of view in the reference point cloud; select a point whose distance between the point in the first field of view and the second nearest neighbor point is within a predetermined range. Determined as the first reference point, wherein the second nearest neighbor point is the point closest to each point in the first field of view determined from the points in the second field of view;

   Alternatively, the first reference point and the second reference point are determined according to the boundary fitting function of the overlapping area in the reference point cloud.
10. The method of claim 9, further comprising:

    Determine the serial numbers of the first reference point and the second reference point;

    The first candidate point and the second candidate point are determined according to the serial numbers of the first reference point and the second reference point.
11. The method of claim 9, further comprising: obtaining a region of interest (ROI) in which the first reference point and the second reference point are determined.
12. A device for determining the layering of lidar point clouds, wherein the lidar includes a first field of view and a second field of view, and there is an overlapping area between the first field of view and the second field of view; The devices include:

    An acquisition module, configured to: acquire a first candidate point of the point cloud of the lidar, where the first candidate point is a point included in the first field of view and located in the overlapping area; Obtain a second candidate point of the LiDAR point cloud, wherein the second candidate point is a point included in the second field of view and located in the overlapping area;

    A determination module configured to determine whether the point cloud is stratified according to a comparison result between the predetermined characteristics of the first candidate point and the predetermined characteristics of the second candidate point.
13. The device according to claim 12, wherein the device is configured to determine that the laser radar has at least one of the following abnormalities in response to determining that stratification occurs in the point cloud: displacement of the laser, displacement of the photodetector, Abnormal behavior of MEMS galvanometer and abnormal internal clock.
14. The device according to claim 12, wherein the predetermined characteristics include one or more of the following: a distance measurement value of the candidate point; an intensity of the candidate point; and a normal vector of the candidate point.
15. The apparatus of claim 14, wherein when the predetermined feature includes a ranging value of an alternative point, the determining module is configured to respond to a mean sum of the ranging values of the first alternative point If the difference between the mean values of the ranging values of the second candidate point is within the first threshold range, it is determined that the point cloud is not stratified; and/or in response to the first candidate point If the difference between the mean value of the ranging values and the mean value of the ranging values of the second candidate point is outside the first threshold range, it is determined that the point cloud is stratified.
16. The apparatus of claim 15, wherein the determining module is configured to respond to an average of the ranging values of the first alternative point and an average of the ranging values of the second alternative point The difference between them is within the first threshold range, and the difference between the standard deviation of the distance measurement value of the first alternative point and the standard deviation of the distance measurement value of the second alternative point If the value is within the second threshold range, it is determined that the point cloud is not stratified.
17. The apparatus according to claim 14, wherein when the predetermined characteristic includes the intensity of candidate points, the determining module is configured to respond to the intensity of all the first candidate points and all the nearest neighbor points. If the mean value of the difference between the intensities is within the third threshold range, it is determined that the point cloud has not been stratified, wherein the nearest neighbor point is from the second alternative point in the second field of view. The determined point closest to each of the first alternative points in the first field of view; and/or, in response to the intensity of all the first alternative points and the nearest neighbor point The mean value of the difference between the intensities is outside the third threshold range, it is determined that the point cloud is stratified, wherein the nearest neighbor point is the second candidate point from the second field of view The point closest to each of the first candidate points in the first field of view is determined.
18. The apparatus of claim 14, wherein when the predetermined feature includes a normal vector of an alternative point, the determining module is configured to respond to a first normal vector determined based on the first alternative point and the first normal vector determined based on the first alternative point. between the second normal vectors determined by the second alternative points angle is within the fourth threshold range, it is determined that the point cloud has not been layered; and/or, in response to the first normal vector determined based on the first alternative point and the third normal vector determined based on the second alternative point. If the angle between the two normal vectors is outside the fourth threshold range, it is determined that the point cloud is stratified.
19. The device according to any one of claims 12 to 18, wherein the determining module is configured to determine that the point cloud occurs in response to the comparison result of the N frame point clouds in the M frame point cloud being a predetermined comparison result. Stratification, wherein said M and said N are integers greater than 0, M≥N.
20. The device according to any one of claims 12-18, wherein the device is also used for:

    Generate a reference point cloud according to the calibration angle and fixed distance of the lidar;

    Determine a first nearest neighbor point in the first field of view in the reference point cloud; select a point whose distance between a point in the second field of view and the first nearest neighbor point is within a predetermined range Determined as a second reference point, wherein the first nearest neighbor point is the point closest to each point in the second field of view determined from the points in the first field of view; Determine a second nearest neighbor point in the second field of view in the reference point cloud; select a point whose distance between the point in the first field of view and the second nearest neighbor point is within a predetermined range. Determined as the first reference point, wherein the second nearest neighbor point is the point closest to each point in the first field of view determined from the points in the second field of view;

    Alternatively, the first reference point and the second reference point are determined according to the boundary fitting function of the overlapping area in the reference point cloud.
21. The device according to claim 20, wherein the device is further used to determine the serial numbers of the first reference point and the second reference point; according to the serial numbers of the first reference point and the second reference point The first alternative point and the second alternative point are determined.
22. The device of claim 20, wherein the device is further configured to acquire a region of interest (ROI) in which the first reference point and the second reference point are determined.
23. A device for determining lidar point cloud layering, including:

    Memory, which stores computer-executable instructions;

    A processor, connected to the memory, configured to implement the method provided in any one of claims 1 to 11 by executing the computer-executable instructions.
24. A computer storage medium, wherein the computer storage medium stores computer-executable instructions; after the computer-executable instructions are executed by a processor, the method as provided in any one of claims 1 to 11 can be implemented.