CN118962639A

CN118962639A - Point cloud noise point identification method and device, laser radar and computer readable storage medium

Info

Publication number: CN118962639A
Application number: CN202310547676.1A
Authority: CN
Inventors: 鲁沛昕
Original assignee: Suteng Innovation Technology Co Ltd
Current assignee: Suteng Innovation Technology Co Ltd
Filing date: 2023-05-15
Publication date: 2024-11-15

Abstract

The application provides a point cloud noise point identification method, a point cloud noise point identification device, a laser radar and a computer readable storage medium, wherein the method comprises the following steps: acquiring a current point cloud point in an initial point cloud, wherein the initial point cloud comprises a plurality of point cloud points detected in a preset time length; when the distance and reflectivity of the current point cloud point meet preset conditions, selecting a target point cloud from the initial point Yun Zhongshai, wherein the target point cloud comprises the current point cloud point and at least part of point cloud points detected in a preset time range before and/or after the current point cloud point; selecting an effective point cloud from the target point Yun Zhongshai, wherein the effective point cloud comprises effective point cloud points with the distance and the reflectivity meeting the preset conditions; and determining whether the current point cloud point is a noise point according to at least one attribute value of the effective point cloud. The application can reduce the influence of noise on the detection of the laser radar.

Description

Point cloud noise point identification method and device, laser radar and computer readable storage medium

Technical Field

The application relates to the field of laser detection, in particular to a point cloud noise point identification method, a point cloud noise point identification device, a laser radar and a computer readable storage medium.

Background

The laser radar is an active remote sensing device for detection by adopting a photoelectric technology, combines the photoelectric detection technology and a laser technology, and is an advanced detection mode by adopting laser as a detection light source. The laser radar comprises a transmitting module, a scanning module, a receiving module and a data processing module, wherein the transmitting module is used for transmitting detection signals to a target, the scanning module is used for changing the transmission defense line of the detection signals, the receiving module is used for receiving echo signals of the detection signals, and the data processing module is used for processing the echo signals, so that information such as the distance, the reflectivity, the speed and the size of the detection target is obtained. The laser radar is widely applied to the fields of automatic driving, vehicle-road coordination, logistics vehicles, robots, public intelligent transportation and the like.

However, in the actual detection process, when the transmitted detection signal encounters interference features such as rain, fog and dust, an echo signal can be generated, the echo signal is received by the receiving module and detected by the data processing module, noise points of the rain, fog and dust can be generated at corresponding positions, and if the noise point features are not processed, the application of the laser radar in application scenes such as automatic driving can be affected.

Disclosure of Invention

The application provides a point cloud noise point identification method, a point cloud noise point identification device, a laser radar and a computer readable storage medium, which can reduce the influence of noise points on detection of the laser radar.

In a first aspect, the method for identifying point cloud noise includes:

acquiring a current point cloud point in an initial point cloud, wherein the initial point cloud comprises a plurality of point cloud points detected in a preset time length;

when the distance and reflectivity of the current point cloud point meet preset conditions, selecting a target point cloud from the initial point Yun Zhongshai, wherein the target point cloud comprises the current point cloud point and at least part of point cloud points detected in a preset time range before and/or after the current point cloud point;

selecting an effective point cloud from the target point Yun Zhongshai, wherein the effective point cloud comprises effective point cloud points with the distance and the reflectivity meeting the preset conditions;

And determining whether the current point cloud point is a noise point according to at least one attribute value of the effective point cloud.

Optionally, when the distance of the current point cloud point is not greater than a first distance threshold value and the reflectivity of the current point cloud point is not greater than a first reflectivity threshold value, determining that the distance and the reflectivity of the current point cloud point meet preset conditions.

Optionally, the at least one attribute value of the valid point cloud includes at least one of:

The number distribution of the effective point clouds,

The distance distribution of the effective point cloud,

And the height distribution of the effective point cloud.

Optionally, the selecting a target point cloud from the initial point Yun Zhongshai includes:

And determining at least one point cloud set from the initial point cloud, wherein the target point cloud comprises the at least one point cloud set, wherein each point cloud point in the same point cloud set is positioned in the same row or the same column, the difference between the maximum pitch angle and the minimum pitch angle in each point cloud point in the same row is smaller than a first preset angle, and the difference between the maximum azimuth angle and the minimum azimuth angle in each point cloud point in the same column is smaller than a second preset angle.

Optionally, the at least one point cloud includes:

A point cloud set including a current point cloud point and a first preset number of point cloud points adjacent to the current point cloud point in a row where the current point cloud point is located; and

The point cloud set comprises a current point cloud point and a second preset numerical point cloud point adjacent to the current point cloud point in a column where the current point cloud point is located.

Optionally, the initial point cloud is scanned by a lidar in a first scanning mode, and the at least one point cloud is determined from the initial point cloud according to the first scanning mode.

Optionally, when the first scanning mode makes the arrangement density of the point cloud points in the same row in the initial point cloud be greater than the arrangement density of the point cloud points in the same column, the at least one point cloud includes at least two point clouds corresponding to the row respectively, or includes at least two point clouds corresponding to the row and the column respectively, and the number of the point clouds corresponding to the row is greater than the number of the point clouds corresponding to the column; or alternatively

When the first scanning mode enables the arrangement density of the point cloud points in the same row in the initial point cloud to be smaller than the arrangement density of the point cloud points in the same column, the at least one point cloud comprises at least two point clouds respectively corresponding to columns or comprises at least two point clouds respectively corresponding to rows and columns, and the number of the point clouds of the corresponding columns is larger than that of the point clouds of the corresponding rows.

Optionally, the determining, according to at least one attribute value of the valid point cloud, whether the current point cloud point is a noise point includes:

Determining whether the current point cloud point is a noise point according to at least one of the following:

the number of the effective point cloud points in each point cloud set,

Continuity of at least one attribute value of the valid point cloud points within each point cloud set,

Similarity of at least one attribute value between the active point clouds of each point cloud.

determining whether the current point cloud point is a noise point according to at least one of the following four items:

whether the number of the effective point cloud points in each point cloud set is respectively larger than a preset number threshold value,

Whether the continuity of the distance distribution of the effective point cloud points in each point cloud set is respectively smaller than a preset continuity threshold value,

Whether the similarity of the distance distribution of the effective point cloud points among the point cloud sets is larger than a first preset similarity threshold value,

And whether the similarity of the height distribution of the effective point cloud points among the point cloud sets is larger than a second preset similarity threshold value or not.

Optionally, determining whether the current point cloud point is a noise point according to the weighted summation result of at least two of the four terms.

Optionally, the continuity of the distance distribution of the effective point cloud points within the point cloud is calculated according to at least one of the following two terms:

The fitting error of the distances of the effective point cloud points in each point cloud set is calculated according to the difference between the distances of the effective point cloud points in each point cloud set and the corresponding fitting distances, the fitting distance corresponding to the effective point cloud points is the fitting distance obtained according to the positions of the effective point cloud points and a fitting function, and the fitting function is a function between the distances obtained by fitting the distances and the positions of the effective point cloud points in each point cloud set;

And/or the number of the groups of groups,

In each point cloud set, the similarity of the number of forward changes, the number of reverse changes and the number of weak changes of the attribute values of two adjacent effective point cloud points,

Wherein, when the fitting error is larger, the continuity of the distance distribution of the effective point cloud points in the point cloud set is smaller;

Wherein the smaller the continuity of the distance distribution of the effective point cloud within the point cloud set, when the number of occurrences of forward direction change, reverse direction change, and weak change is more similar.

Optionally, the similarity of the distance distribution of the effective point cloud points between the point cloud sets is determined according to at least one of the following two:

A difference between a distance of the effective point cloud point of each point cloud and a distance average of the point cloud;

similarity between distance averages of each point cloud.

Optionally, the similarity of the height distribution of the effective point cloud points between the point cloud sets is determined according to at least one of the following two:

a difference between the height of the effective point cloud point within each point cloud and the average of the heights of the point clouds;

similarity between the mean height values of the point clouds.

Optionally, the method further comprises:

Determining candidate point cloud points including point cloud points in which a distance and a reflectivity in the initial point cloud satisfy the preset condition, and determining not the noise point,

And checking whether the candidate point cloud point is the noise point according to the number of the point cloud points which are determined to be the noise point and are positioned in the neighborhood point cloud of the candidate point cloud point.

Optionally, the rechecking whether the candidate point cloud point is the specific noise point according to the number of the point cloud points which are located in the neighborhood point cloud of the candidate point cloud point and are the noise point includes:

And rechecking the candidate point cloud point as the noise point when the number of the point cloud points which are the noise points in the neighborhood point cloud of the candidate point cloud point reaches a second number threshold.

Optionally, the neighborhood point cloud of the candidate point cloud point is a point cloud array including L1 x L2 of the candidate point cloud point, and L1 and L2 are positive integers respectively.

Optionally, the method further comprises:

Removing the noise point from the initial point cloud; or alternatively

The noise point is marked in the initial point cloud.

In a second aspect, the present application provides a point cloud noise point identification device, including:

The acquisition module is used for acquiring the current point cloud point in the initial point cloud, wherein the initial point cloud comprises a plurality of point cloud points detected in a preset time length;

A first screening module, configured to select a target point cloud from the initial point Yun Zhongshai when the distance and reflectivity of the current point cloud point satisfy preset conditions, where the target point cloud includes the current point cloud point and at least part of the point cloud points detected in a preset time range before and/or after the current point cloud point;

The second screening module is configured to select an effective point cloud from the target points Yun Zhongshai, where the effective point cloud includes effective point cloud points whose distance and reflectivity meet the preset conditions;

And the first determining module is used for determining whether the current point cloud point is a noise point according to at least one attribute value of the effective point cloud.

In a third aspect, the present application provides a lidar comprising:

A processor; and

A memory having executable code stored thereon that, when executed by the processor, causes the processor to perform the method of any of the claims.

In a fourth aspect, the present application provides a computer readable storage medium storing executable code which, when executed by a processor of an electronic device, causes the processor to perform the method of any one of the above.

In the embodiment of the application, the characteristics of the distance and the reflectivity of the noise points corresponding to the tiny particle objects (such as rain, fog, dust and the like) in the current detection environment are utilized, the point cloud points with the distance and the reflectivity meeting the preset conditions are used as the current point cloud points, at least part of the detected point cloud points in the preset time range before and/or after the current point cloud points are screened out to be used as effective point clouds, the characteristics of clustering the noise points and the distribution among the attribute values of each noise point are utilized, and whether the current point cloud points are the noise points or not is determined according to at least one attribute value of each effective point cloud point in the effective point clouds, so that the influence of the noise points on the detection of the laser radar is reduced.

Drawings

FIG. 1 is a schematic diagram of one embodiment of a point cloud noise point identification method of the present application;

FIG. 2 is a schematic diagram of one embodiment of a point cloud noise point identification apparatus in the present application;

fig. 3 is a schematic diagram of one embodiment of a lidar of the present application.

Detailed Description

Embodiments of the present application will be described in more detail below with reference to the accompanying drawings. While embodiments of the present application are illustrated in the drawings, it should be understood that the present application may be embodied in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the application to those skilled in the art.

The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. As used in this specification and the appended claims, the singular forms "a," "an," and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It should also be understood that the term "and/or" as used herein refers to and encompasses any or all possible combinations of one or more of the associated listed items.

It should be understood that although the terms "first," "second," "third," etc. may be used herein to describe various information, these information should not be limited by these terms. These terms are only used to distinguish one type of information from another. For example, first information may also be referred to as second information, and similarly, second information may also be referred to as first information, without departing from the scope of the application. Thus, a feature defining "a first" or "a second" may explicitly or implicitly include one or more such feature. In the description of the present application, the meaning of "a plurality" is two or more, unless explicitly defined otherwise.

The laser radar is a radar system for detecting the position, speed and other characteristic quantities of a target by emitting laser beams, and the working principle of the laser radar is that a plurality of laser beams are sequentially emitted by a laser emitter, if the laser beams meet an object, the laser beams are reflected, and a receiver is used for receiving the reflected echoes. The laser radar may be a mechanical laser radar, a semi-solid laser radar, a pure solid laser radar, or the like. The lidar calculates a time difference, which is then the time of flight of the laser beam, from the time the reflected echo was received and the time the laser beam was emitted. The laser radar calculates parameters such as distance, azimuth and the like of an object reflecting the echo according to the flight time, and calculates reflectivity according to the energy of the echo and the energy of the emitted laser beam, so that the external environment is detected.

Lidar generally includes a transceiver module and a processing module. The receiving and transmitting module converts the electric signal into an optical signal and transmits the optical signal, the optical signal reflected by the object is restored into an electric signal, and the processing module calculates relevant information according to the electric signal. Optionally, the laser radar may further include a scanning module (such as a MEMS two-dimensional galvanometer, a one-dimensional galvanometer, a turning mirror, a turning platform, etc., the present application is not limited to a specific form of the scanning module), and is configured to change the emitting direction of the emitted laser beam, so as to implement scanning of the laser beam within a certain field of view. The optical paths of the reflected echo and the outgoing laser light can be coaxial, and the reflected echo returns along the path of the scanning module-transceiver module and is received by the receiver in the transceiver module. The laser radar can be a single-channel scanning or a multi-channel simultaneous or staggered scanning, wherein in the case of multi-channel scanning, the multi-channel scanning can be realized through a plurality of groups of transceiver modules, and light rays (emergent laser and reflected echo) between the channels are isolated from each other.

After the laser radar calculates parameters such as distance, azimuth, reflectivity and the like of the reflected echo, a point cloud point is used for representing an object reflecting the echo. And (5) collecting the point cloud points in a continuous period of time to obtain a frame of point cloud. In general, the time length of each frame of point cloud frame output by the laser radar is the same, that is, the point cloud points collected in the period are output as one frame of point cloud every fixed period. Of course, in some cases, the point cloud frames may also be output at a variable frequency, which is not limiting herein. At least one frame of point cloud acquired by the laser radar can be used for further identifying and sensing the external environment.

However, each echo received by the lidar is not necessarily reflected by an object having a certain volume in the external environment, and may be reflected by a tiny particle object such as dust, rainwater, fog, hail, etc. in the external environment. These tiny particle objects are not detection targets of the lidar, and the presence of point cloud points corresponding to these tiny particle objects in the point cloud may instead affect decisions made based on the identification of the point cloud, such as deciding on the path planning of moving objects (e.g., automobiles, robots, etc.) based on obstacles identified by the point cloud. Therefore, the point cloud points corresponding to these tiny particle objects are regarded as noise points to be identified or removed.

The point cloud noise point identification method of the present application is described below by way of example with reference to fig. 1. As shown in fig. 1, fig. 1 is a schematic diagram of an embodiment of a point cloud noise point identification method of the present application. The point cloud noise point identification method comprises the following steps:

Step S101, a current point cloud point in an initial point cloud is obtained, wherein the initial point cloud comprises a plurality of point cloud points detected in a preset time period.

The initial point cloud in the present application includes a plurality of point cloud points detected within a preset time period, for example, may refer to one frame of point cloud, may refer to point cloud in a part of areas in one frame of point cloud, or may be a set of continuous multi-frame point cloud, which is not limited herein. The initial point cloud may be obtained by a lidar. The execution subject of the point cloud noise point identification method may be the lidar. For example, the laser radar determines which points in the initial point cloud are noise points according to the point cloud noise point identification method in the present embodiment after each acquisition of a point cloud of a partial area of a one-frame point cloud, or each acquisition of a multi-frame point cloud.

For another example, the main body of execution of the point cloud noise point recognition method may be executed by a processing device other than the laser radar, for example, a processing device in a control system in an automobile or a robot. The processing device determines which points in the initial point cloud are noise points by acquiring the initial point cloud obtained by laser radar scanning according to the point cloud noise point identification method in the embodiment.

Step S102, when the distance and reflectivity of the current point cloud point satisfy the preset conditions, selecting a target point cloud from the initial point Yun Zhongshai, where the target point cloud includes the current point cloud point and at least part of the point cloud points detected in a preset time range before and/or after the current point cloud point.

Alternatively, step S102 may be performed with all the point cloud points in the initial point cloud sequentially as the current point cloud points. Or step S102 may be performed sequentially for only the point cloud points in the partial field of view of the initial point cloud as the current point cloud points. For example, in an example where a point cloud is used for a path planning application of a moving object, an obstacle located above the height of the moving object does not affect the path planning of the moving object, then in identifying a noise point, noise point identification may be performed only for a point cloud point located within a field of view below the height of the moving object in the initial point cloud as a current point cloud point.

The distance and the reflectivity of the point cloud point serving as the noise point have certain characteristics, and the method is used as a preliminary screening condition of the noise point by confirming whether the distance and the reflectivity of the current point cloud point meet preset conditions. In one example, when the distance of the current point cloud point is not greater than a first distance threshold and the reflectivity of the current point cloud point is not greater than a first reflectivity threshold, it is determined that the distance and the reflectivity of the current point cloud point satisfy a preset condition. Because the capability of reflecting laser beams by tiny particle objects such as rain, fog, dust and the like is weak, the reflectivity of point cloud points corresponding to the tiny particle objects is low, and because the capability of reflecting laser beams is weak, echoes reflected by far tiny particle objects are too weak to be detected, and therefore the distance and the reflectivity of the point cloud points belonging to noise points in the detected echoes are respectively smaller than a certain value. The first distance threshold and the first reflectivity threshold may be determined by experimental or empirical values. For example, the first distance threshold may be a value between 20cm and 30 cm.

The noise points corresponding to the micro-particle objects such as rain, fog, dust and the like generally appear in a lump, when the distance and the reflectivity of the current point cloud point are determined to meet the preset conditions, the target point cloud is determined from the point clouds around the current point cloud point according to the position relation with the current point cloud point, and whether the current point cloud point is the noise point is determined by utilizing the target point cloud.

And step S103, screening effective point clouds with the distance and the reflectivity meeting the preset conditions from the target point clouds.

The noise points corresponding to the micro-particle objects such as rain, fog, dust and the like generally appear in a lump, and the characteristic changes are uneven and random. Therefore, after the target point cloud is determined, effective point clouds with the distance and the reflectivity meeting the preset conditions are screened out from the target point cloud, the probability that the effective point clouds are noise points is larger than the probability that the point clouds which do not meet the preset conditions in the target point cloud are noise points, and by analyzing the attribute values of the effective point clouds, whether the current point clouds are noise points or not is determined by analyzing the attribute values of the target point clouds, the complexity can be reduced, and the accuracy can be improved.

Step S104, determining whether the current point cloud point is a noise point according to at least one attribute value of the effective point cloud.

Optionally, the at least one attribute value of the valid point cloud includes at least one of: the quantity of the effective point clouds is distributed, the distance of the effective point clouds is distributed, and the height of the effective point clouds is distributed. For example, the number distribution of the effective point clouds may be whether the number of effective point clouds is greater than a preset threshold. For example, the distance distribution of the effective point cloud may be a continuity and/or similarity of the change in distance of each point cloud point in the effective point cloud. As another example, the continuity and/or similarity of the change in height of each cloud point in the effective point cloud. The distribution probability of the distance and/or the height of noise points generated by general tiny particle objects (such as rain, fog and dust) on local is random, and the distribution probability is continuous or regular; and generally has the characteristic of agglomeration, so that the distance or the height generally has a certain similarity. And whether the current point cloud point is a noise point with high probability can be judged by analyzing the attribute values of the effective point cloud point.

In the embodiment of the application, by utilizing the characteristics of the distance and the reflectivity of the noise points corresponding to the tiny particle objects in the current detection environment, whether the current point cloud points are noise points or not is determined by taking the point cloud points with the distance and the reflectivity meeting the preset conditions as the current point cloud points and screening out at least part of the point cloud points detected in the preset time range before and/or after the current point cloud points as effective point clouds, and by utilizing the characteristics of clustering the noise points and the distribution among the attribute values of each noise point, the influence of the noise points on the detection of the laser radar is reduced.

In one example, since noise points corresponding to tiny particle objects such as rain, fog, dust and the like generally appear in a cluster shape, and feature changes in the transverse direction and the longitudinal direction are uneven and random, the target attribute values of point cloud points in the target point cloud can be analyzed in the transverse direction and/or the longitudinal direction. The following describes how to determine a point cloud for lateral analysis and/or longitudinal analysis and how to determine whether a current point cloud point is a noisy point based on the point cloud.

In step S102, when a target point cloud is selected from the initial points Yun Zhongshai, at least one point cloud is determined as the target point cloud according to the spatial arrangement of the point cloud points or according to the scanning order of the lidar. Wherein, each point cloud point in the same point cloud set is located in the same row or the same column, wherein, the difference between the maximum pitch angle and the minimum pitch angle in each point cloud point in the same row is smaller than a first preset angle, and the difference between the maximum azimuth angle and the minimum azimuth angle in each point cloud point in the same column is smaller than a second preset angle.

For example, taking a point cloud set as an example of a row of point clouds, taking a current point cloud point as a starting point, or as an ending point, or as an intermediate point, determining a preset number of point cloud points in the same row, which are adjacent to the current point cloud point, as the current row. The current line may be used as the first line, the last line, or the middle line, at least one line of point cloud points adjacent to the current line is determined, and the current line forms a point cloud array as the target point cloud. The method of selecting the target point cloud using the point cloud as a list of point clouds is similar to the method of selecting the target point cloud using the point cloud as a row of point clouds.

Further optionally, the target point cloud includes at least one point cloud of a corresponding row and at least one point cloud of a corresponding column. Specifically, the target point cloud includes: a current point cloud point and a point cloud set of a first preset number of point cloud points adjacent to the current point cloud point in a row where the current point cloud point is located; and a point cloud set comprising a current point cloud point and a second preset number of point cloud points adjacent to the current point cloud point in a column where the current point cloud point is located. The first preset value and the second preset value may be the same or different.

Different scanning modes of the laser radar may cause the point cloud to have different distributions, and in some examples, the scanning mode of the laser radar is to sequentially scan along one row and then move to the next row to sequentially scan, or sequentially scan along one column and then move to the next column to sequentially scan, or sequentially scan along circles or ellipses with different diameters, or sequentially scan along a spiral line, and so on. In the example, a row of point clouds is defined through pitch angles of the cloud points of each point, wherein the pitch angles of the cloud points of each point in the row of point clouds are similar, and the difference between the maximum pitch angle and the minimum pitch angle is smaller than a first preset angle; and defining a row of point clouds by azimuth angles of all the point clouds, wherein azimuth angles of all the point clouds in a row of point clouds are similar, and the difference between the maximum azimuth angle and the minimum azimuth angle is smaller than a second preset angle. Wherein the first preset angle and the second preset angle may be the same or different.

Optionally, the initial point cloud is scanned by a lidar in a first scanning mode, and the at least one point cloud is determined from the initial point cloud according to the first scanning mode. In one example, when the first scanning mode makes the arrangement density of the point cloud points in the same row in the initial point cloud be greater than the arrangement density of the point cloud points in the same column, the at least one point cloud includes at least two point clouds respectively corresponding to the row or includes at least two point clouds respectively corresponding to the row and the column, and the number of the point clouds of the corresponding row is greater than the number of the point clouds of the corresponding column. In one example, when the first scanning mode makes the arrangement density of the point cloud points in the same row in the initial point cloud smaller than the arrangement density of the point cloud points in the same column, the at least one point cloud includes at least two point clouds respectively corresponding to columns or includes at least two point clouds respectively corresponding to rows and columns, and the number of the point clouds of the corresponding columns is greater than the number of the point clouds of the corresponding rows.

For example, in practical application, when the scanning mode of the laser radar is to sequentially scan along a row and then move to the next row to sequentially scan, the arrangement density of the point cloud points in the same row is greater than the arrangement density of the point cloud points in the same column. When the target point cloud is selected from the initial point Yun Zhongshai, the point cloud points obtained by N continuous scans of the current point cloud point in the current row of the point cloud point where the current point cloud point is located can be added as one point cloud set to the target point cloud, or M point cloud points adjacent to the N point cloud points in each row of at least one row of the point cloud adjacent to the current row of the point cloud point can be added as one point cloud set to the target point cloud. Wherein N and M are positive integers, and N and M can be the same or different. Optionally, in the current column of point clouds where the current point is located, point clouds obtained by X continuous scans including the current point cloud point may be added as a point cloud set to the target point cloud, or Y point clouds adjacent to the X point clouds may be added as a point cloud set to the target point cloud in each column of at least one column of point clouds adjacent to the current column. Wherein X and Y are positive integers, X and Y can be the same or different, and X and M can be the same or different.

In a specific example, N point cloud points including the current point cloud point in the current row, N point cloud points closest to the N point cloud points in the two row point clouds scanned after the current row, and N point cloud points including the current point cloud point in the current column may be added to the target point cloud as four point clouds, a point cloud array of 3*N and a point cloud array of X1 in total.

In one example, when the first scanning mode makes the arrangement density of the point cloud points in the same row in the initial point cloud smaller than the arrangement density of the point cloud points in the same column, the at least one point cloud includes at least two point clouds respectively corresponding to columns or includes at least two point clouds respectively corresponding to rows and columns, and the number of the point clouds of the corresponding columns is greater than the number of the point clouds of the corresponding rows.

When the scanning mode of the laser radar is to sequentially scan along a column and then move to the next column to continue scanning in sequence, the arrangement density of the point cloud points in the same column is greater than that in the same row. When the target point cloud is selected from the initial point Yun Zhongshai, the point cloud points obtained by N continuous scans of the current point cloud point in the current column of the point cloud where the current point cloud point is located may be added as one point cloud set to the target point cloud, or M point cloud points adjacent to the N point cloud points may be added as one point cloud set to the target point cloud in each column of at least one column of the point cloud adjacent to the current column. Optionally, in the current row of point clouds where the current point is located, point clouds obtained by X continuous scanning including the current point cloud point may be added as a point cloud set to the target point cloud, or Y point cloud points adjacent to the X point clouds may be added as a point cloud set to the target point cloud in each row of at least one row of adjacent point clouds of the current row.

In a specific example, X point cloud points including the current point cloud point in the current column, and X point cloud points closest to the X point cloud points in the two column point clouds scanned after the current column, and N point cloud points including the current point cloud point in the current row may be added as four point clouds, a point cloud array of 3*X and a point cloud array of n×1 in total, to the target point cloud.

In the example where the target point cloud is a plurality of point clouds, after all the effective point clouds are screened out from the point clouds, the effective point clouds may be analyzed in units of point clouds. In some examples, when determining whether the current point cloud point is a noise point according to at least one attribute value of the valid point cloud, the determining whether the current point cloud point is a noise point may be performed according to at least one of the following four terms: (1) whether the number of the effective point cloud points in each point cloud set is respectively larger than a preset number threshold, (2) whether the continuity of the distance distribution of the effective point cloud in each point cloud set is respectively smaller than a preset continuity threshold, (3) whether the similarity of the distance distribution of the effective point cloud between each point cloud sets is larger than a first preset similarity threshold, and (4) whether the similarity of the height distribution of the effective point cloud between each point cloud set is larger than a second preset similarity threshold. Further alternatively, it may be determined whether the current point cloud point is a noise point according to a weighted sum result of at least two of the four items.

In one example, the continuity of the distance distribution of the effective point cloud within each point cloud in item (2) may be obtained from at least one of: (a) Whether the fitting error of the distances of the effective point cloud points in each point cloud set is larger than a preset fitting error threshold value, and (b) the similarity of the number of forward changes, the number of reverse changes and the number of weak changes of the attribute values of two adjacent effective point cloud points in each point cloud set. For example, when the fitting error of the distances between the effective point cloud points in one point cloud set is greater than a preset fitting error threshold, and/or the similarity of the number of forward changes, the number of reverse changes and the number of weak changes of the attribute values of two adjacent effective point cloud points in the point cloud set is greater than a third preset similarity threshold, the continuity of the distance distribution of the effective point cloud in the point cloud set is less than a preset continuity threshold.

In the step (a), the fitting error is calculated according to the difference between the distance between each point cloud point of the effective point cloud in the current point cloud set and the corresponding fitting distance, the fitting distance corresponding to the point cloud point is the fitting distance obtained according to the position of the point cloud point and a fitting function, and the fitting function is a function between the distance obtained according to the fitting of the distance between each point cloud point of the effective point cloud in the current point cloud set and the position. The greater the fitting error, the less the continuity of the distance distribution within the point cloud of the effective point cloud.

In a specific example, taking a target point cloud as an example of three point clouds corresponding to three rows of point clouds respectively, an array is formed by using effective point clouds in each point cloud, and the position of each effective point cloud point in the array corresponds to the position determination in the row. Thus, according to the position of each point cloud point in the array and the distance of the point cloud point, a function y=f (x) between the distance and the position can be fitted, wherein x represents the position of the point cloud point in the array, and y represents the distance of the point cloud point. The function to be fitted may be a quadratic function, a linear function, an inverse proportional function, a multiple function, or the like, and is not limited thereto. The difference between the actual distance of each point cloud point and the fitting distance calculated from the function y=f (x) is then calculated. And calculating fitting errors according to the corresponding differences of the cloud points of the effective points. The larger the fitting error is, the more irregular the distance arrangement of the cloud points of each point is, the smaller the continuity of the distance distribution of the effective point cloud in the point cloud set is, the smaller the probability that each cloud point corresponds to the surface of the obstacle object is, and the larger the probability that each cloud point has noise points with random distances is.

In the step (b), the forward change, the reverse change and the weak change in one point cloud set refer to that each effective point cloud point in the point cloud set is arranged in a row or a column of point clouds, and when the difference between the distance of the next point cloud point and the distance of the previous point cloud point is larger than a preset distance value, the forward change occurs once; the difference between the distance of the previous point cloud point and the distance of the next point cloud point is larger than a preset distance value, a reverse change occurs, and when the absolute value of the difference between the distance of the previous point cloud point and the distance of the next point cloud point is not larger than the preset distance value, a weak change occurs. The specific value of the preset distance value can be determined according to an empirical value or an experiment.

In one example, the similarity of the distance distribution of the effective point clouds between the point clouds in item (3) may be determined according to at least one of: (a) A difference between a distance of the effective point cloud point of each point cloud and a distance average of the point cloud; (b) similarity between distance averages of point clouds. Specifically, when the distance errors calculated by the differences between the distances of the effective point cloud points in each point cloud set and the distance average value of the point cloud set are smaller than the preset distance errors, and the similarity of the distance average value between the point cloud sets is larger than the preset similarity, determining that the similarity of the distance distribution of the effective point cloud points between the point cloud sets is larger than a first preset similarity threshold.

In one example, the similarity of the height distribution of the effective point cloud between the point clouds in item (4) may be determined according to at least one of: (a) A difference between the height of the effective point cloud point within each point cloud and the average of the heights of the point clouds; (b) similarity between the mean height values of the point clouds. Specifically, when the height error calculated by the difference between the height of each effective point cloud point in each point cloud set and the average value of the heights of the point cloud sets is smaller than the preset height error, and the similarity of the average value of the heights of the point cloud sets is larger than the preset similarity, determining that the similarity of the height distribution of the effective point cloud points between the point cloud sets is larger than a second preset similarity threshold.

Since the noise points corresponding to the tiny particle objects generally have the characteristic of agglomeration, whether the distances or the heights of the effective point cloud points in the whole are similar or not is analyzed through the (3) item and the (4) item, and whether the current point cloud points are the noise points or not is judged.

Optionally, after all the point cloud points serving as the noise points are determined according to the point cloud noise point identification method by taking the point cloud points in the initial point cloud as the current point cloud points in sequence, whether the point cloud points which are not the noise points are the noise points or not can be checked in the effective point cloud by utilizing the characteristic that the noise points corresponding to the tiny particle objects are clustered in the whole. Specifically, the point cloud noise point identification method in the application further comprises the following steps: determining candidate point cloud points, wherein the candidate point cloud points comprise point cloud points which are not the noise points, and the distance and the reflectivity of the candidate point cloud points in the initial point cloud meet the preset conditions; and checking whether the candidate point cloud point is the noise point according to the number of the point cloud points which are determined to be the noise point and are positioned in the neighborhood point cloud of the candidate point cloud point.

The method for determining the neighborhood point cloud of the candidate point cloud point is various, for example, a point cloud array including L1 x L2 of the candidate point cloud point may be used as the neighborhood point cloud, and L1 and L2 are positive integers respectively. The candidate point cloud point may be a point cloud at any one location (e.g., at or near the center) of the point cloud array. The condition of checking whether the candidate point cloud point is the noise point may be various, for example, when the number of point cloud points which are the noise points in the neighborhood point cloud of the candidate point cloud point reaches a second number threshold, the candidate point cloud point is checked as the noise point. Optionally, the distances and the reflectivities in the initial point cloud meet the preset conditions in sequence, and the point cloud points which are not the noise points are determined to serve as the candidate point cloud points for rechecking.

After the noise point is determined from the initial point cloud, optionally, the point cloud point determined as the noise point may be marked, or the point cloud point determined as the noise point may be directly removed from the initial point cloud. Alternatively, when marking the noise, the point cloud point that is determined to be the noise during the initial judgment (i.e., the noise is determined to be the noise according to steps S101 to S104) and the point cloud point that is determined to be the noise during the review judgment may be marked in the same manner, or may be marked in different manners, for example, the invalid value of the point cloud point that is determined to be the noise during the initial judgment is marked as 0, and the invalid value of the point cloud point that is determined to be the noise during the review judgment is marked as 1.

The embodiment of the application also provides a device for identifying the point cloud noise point. As shown in fig. 2, fig. 2 is a schematic diagram of an embodiment of a point cloud noise point identification device in an embodiment of the present application. The laser ranging apparatus 200 includes:

An obtaining module 201, configured to obtain a current point cloud point in an initial point cloud, where the initial point cloud includes a plurality of point cloud points detected within a preset duration;

A first screening module 202, configured to select a target point cloud from the initial point Yun Zhongshai when the distance and reflectivity of the current point cloud point satisfy a preset condition, where the target point cloud includes the current point cloud point and at least part of the point cloud points detected in a preset time range before and/or after the current point cloud point;

a second screening module 203, configured to select an effective point cloud from the target points Yun Zhongshai, where the effective point cloud includes effective point cloud points whose distance and reflectivity meet the preset conditions;

The first determining module 204 is configured to determine whether the current point cloud point is a noise point according to at least one attribute value of the valid point cloud.

The number distribution of the effective point clouds,

The distance distribution of the effective point cloud,

And the height distribution of the effective point cloud.

Optionally, the first screening module 202 is specifically configured to, when selecting the target point cloud from the initial point Yun Zhongshai:

Optionally, the at least one point cloud includes:

Optionally, when determining whether the current point cloud point is a noise point according to at least one attribute value of the valid point cloud, the first determining module 204 is specifically configured to:

the number of the effective point cloud points in each point cloud set,

And/or the number of the groups of groups,

similarity between distance averages of each point cloud.

similarity between the mean height values of the point clouds.

Optionally, the apparatus further comprises:

a second determining module for determining candidate point cloud points including point cloud points whose distance and reflectivity satisfy the preset condition in the initial point cloud and which are not the noise points,

And the rechecking module is used for rechecking whether the candidate point cloud point is the noise point according to the number of the point cloud points which are determined to be the noise point and are positioned in the neighborhood point cloud of the candidate point cloud point.

Optionally, the rechecking module is specifically configured to, when rechecking whether the candidate point cloud point is the specific noise point according to the number of point cloud points that are located in the neighborhood point cloud of the candidate point cloud point and are the noise point:

Optionally, the apparatus further comprises:

The rejecting module is used for rejecting the noise point from the initial point cloud; or alternatively

And the marking module is used for marking the noise point in the initial point cloud.

The point cloud noise point identification device in the embodiment of the application utilizes the characteristics of the distance and the reflectivity of the noise points corresponding to tiny particle objects (such as rain, fog, dust and the like) in the current detection environment, and determines whether the current point cloud point is a noise point or not by taking the point cloud point with the distance and the reflectivity meeting preset conditions as the current point cloud point and screening out at least part of the detected point cloud points in a preset time range before and/or after the current point cloud point as effective point cloud, and utilizes the characteristics of clustering of the noise points and distribution among attribute values of each noise point to determine whether the current point cloud point is the noise point or not according to at least one attribute value of each effective point cloud point in the effective point cloud so as to reduce the influence of the noise point on the detection of a laser radar.

The embodiment of the application also provides a laser radar, as shown in fig. 3, and fig. 3 is a schematic diagram of one embodiment of the laser radar in the embodiment of the application. The lidar 30 includes a memory 31 and a processor 32.

The Processor 32 may be a central processing unit (Central Processing Unit, CPU), but may also be other general purpose processors, digital signal processors (DIGITAL SIGNAL Processor, DSP), application SPECIFIC INTEGRATED Circuit (ASIC), field-Programmable gate array (Field-Programmable GATE ARRAY, FPGA) or other Programmable logic device, discrete gate or transistor logic device, discrete hardware components, or the like. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.

The memory 31 may include various types of storage units, such as system memory, read Only Memory (ROM), and persistent storage. Where the ROM may store static data or instructions that are required by the processor 32 or other modules of the computer. The persistent storage may be a readable and writable storage. The persistent storage may be a non-volatile memory device that does not lose stored instructions and data even after the computer is powered down. In some embodiments, the persistent storage device employs a mass storage device (e.g., magnetic or optical disk, flash memory) as the persistent storage device. In other embodiments, the persistent storage may be a removable storage device (e.g., diskette, optical drive). The system memory may be a read-write memory device or a volatile read-write memory device, such as dynamic random access memory. The system memory may store instructions and data that are required by some or all of the processors at runtime. Furthermore, memory 61 may comprise any combination of computer-readable storage media including various types of semiconductor memory chips (e.g., DRAM, SRAM, SDRAM, flash memory, programmable read-only memory), magnetic disks, and/or optical disks may also be employed. In some embodiments, memory 510 may include a readable and/or writable removable storage device, such as a Compact Disc (CD), a digital versatile disc read only (e.g., DVD-ROM, dual layer DVD-ROM), a blu-ray read only disc, an ultra-dense disc, a flash memory card (e.g., SD card, min SD card, micro-SD card, etc.), a magnetic floppy disk, and the like. The computer readable storage medium does not contain a carrier wave or an instantaneous electronic signal transmitted by wireless or wired transmission.

The memory 31 has stored thereon executable code that, when processed by the processor 32, causes the processor 32 to perform some or all of the methods described above.

Furthermore, the method according to the application may also be implemented as a computer program or computer program product comprising computer program code instructions for performing part or all of the steps of the above-described method of the application.

Or the application may also be embodied as a computer-readable storage medium (or non-transitory machine-readable storage medium or machine-readable storage medium) having stored thereon executable code (or a computer program or computer instruction code) which, when executed by a processor of an electronic device (or server, etc.), causes the processor to perform some or all of the steps of the above-described method according to the application.

The foregoing description of embodiments of the application has been presented for purposes of illustration and description, and is not intended to be exhaustive or limited to the embodiments disclosed. Many modifications and variations will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the various embodiments described. The terminology used herein was chosen in order to best explain the principles of the embodiments, the practical application, or the improvement of technology in the marketplace, or to enable others of ordinary skill in the art to understand the embodiments disclosed herein.

Claims

1. The point cloud noise point identification method is characterized by comprising the following steps of:

2. The method of claim 1, wherein the distance and the reflectivity of the current cloud point are determined to satisfy a preset condition when the distance of the current cloud point is not greater than a first distance threshold and the reflectivity of the current cloud point is not greater than a first reflectivity threshold.

3. The method of claim 1, wherein the at least one attribute value of the valid point cloud comprises at least one of:

The number distribution of the effective point clouds,

The distance distribution of the effective point cloud,

And the height distribution of the effective point cloud.

4. A method according to any one of claims 1 to 3, wherein said selecting a target point cloud from said initial point Yun Zhongshai comprises:

5. The method of claim 4, wherein the at least one point cloud comprises:

6. The method of claim 4, wherein the initial point cloud is scanned by a lidar in a first scanning mode, and wherein the at least one point cloud is determined from the initial point cloud according to the first scanning mode.

7. The method of claim 6, wherein when the first scanning pattern is such that the arrangement density of the point cloud points in the same row in the initial point cloud is greater than the arrangement density of the point cloud points in the same column, the at least one point cloud includes at least two point clouds of respective corresponding rows or includes at least two point clouds of respective corresponding rows and columns, and the number of point clouds of the corresponding rows is greater than the number of point clouds of the corresponding columns; or alternatively

8. A point cloud noise point identification device, characterized by comprising:

9. A laser radar which comprises a laser beam source, characterized by comprising the following steps:

A processor; and

A memory having executable code stored thereon, which when executed by the processor causes the processor to perform the method of any of claims 1 to 7.

10. A computer readable storage medium, characterized in that executable code is stored, which when executed by a processor of an electronic device causes the processor to perform the method of any of claims 1 to 7.