CN-122024203-A - Method, equipment, medium and product for extracting point cloud of interest area based on roadside laser radar

CN122024203ACN 122024203 ACN122024203 ACN 122024203ACN-122024203-A

Abstract

The embodiment of the application relates to the field of intelligent traffic and discloses a method, equipment, medium and product for extracting a point cloud of a region of interest based on a roadside laser radar; the method comprises the steps of obtaining three-dimensional point cloud data collected by a road side laser radar, traversing each target point in the three-dimensional point cloud data, calculating mapping pixel coordinates of the target points in a pre-built region-of-interest configuration diagram according to preset resolution parameters and perception range parameters, wherein the region-of-interest configuration diagram is a two-dimensional aerial view diagram, different pixel attribute values represent regions of interest and non-regions of interest, inquiring pixel attribute values corresponding to the mapping pixel coordinates in the region-of-interest configuration diagram, eliminating the target points from the three-dimensional point cloud data if the pixel attribute values correspond to the non-regions of interest, and reserving the target points as the point clouds of the regions of interest if the pixel attribute values correspond to the regions of interest.

Inventors

MA YUAN
XUAN ZHIYUAN
YANG XUAN
ZHANG LIJUAN

Assignees

云控智行科技有限公司

Dates

Publication Date: 20260512
Application Date: 20260119

Claims (10)

1. The method for extracting the point cloud of the region of interest based on the roadside laser radar is characterized by comprising the following steps: Acquiring three-dimensional point cloud data acquired by a roadside laser radar; Traversing each target point in the three-dimensional point cloud data, and calculating mapping pixel coordinates of the target point in a pre-constructed region-of-interest configuration diagram according to preset resolution parameters and perception range parameters, wherein the region-of-interest configuration diagram is a two-dimensional aerial view diagram, and different pixel attribute values represent regions of interest and non-regions of interest; inquiring a pixel attribute value corresponding to the mapped pixel coordinate in the region of interest configuration diagram; And if the pixel attribute value corresponds to the region of interest, reserving the target point as the point cloud of the region of interest.
2. The method according to claim 1, wherein the method for constructing the region of interest map specifically comprises: collecting continuous multi-frame original point cloud data of the road side laser radar in a preset history period; performing background filtering on the original point cloud data, and extracting point cloud data of dynamic traffic participants; Projecting the point cloud data of the dynamic traffic participants to a two-dimensional aerial view coordinate system, and counting the accumulated number or the occurrence frequency of the point clouds at each pixel position to generate a track thermodynamic diagram; And carrying out binarization processing on the track thermodynamic diagram according to a preset density threshold value, marking a region higher than the density threshold value as a region of interest, and marking the rest regions as non-regions of interest so as to generate the region of interest configuration diagram.
3. The method of claim 2, wherein after the generating the region of interest configuration map, the method further comprises: Performing morphological closing operation on the binarized image to fill discontinuous cavities in the region of interest; and/or performing morphological dilation operation on the binarized image, and performing flaring on the edge of the region of interest to form the region of interest containing the safety margin.
4. The method of claim 1, wherein if the pixel attribute value corresponds to a region of interest, then retaining the target point as a region of interest point cloud comprises: When the pixel attribute value corresponds to the region of interest, acquiring a height value of the target point in the three-dimensional point cloud data; Inquiring a corresponding effective height interval of the mapped pixel coordinates in the region of interest configuration diagram; judging whether the height value is in the effective height interval or not; if yes, the target point is reserved as the point cloud of the region of interest, and if not, the target point is removed from the three-dimensional point cloud data.
5. The method of claim 4, wherein the effective height interval is constructed in a manner comprising: When the region of interest configuration diagram is generated, counting the height distribution characteristics of the historical dynamic traffic participant point clouds falling into each pixel coordinate; performing kernel density estimation according to the height distribution characteristics to generate a height probability density function; And determining the minimum effective height and the maximum effective height corresponding to the pixel coordinates according to the main peak distribution range of the height probability density function so as to construct the effective height interval covering all pixels.
6. The method according to any one of claims 1 to 5, wherein the calculating the mapped pixel coordinates of the target point in the region of interest configuration map according to the preset resolution parameter and perception range parameter specifically adopts the following formula: wherein (x, y) is the horizontal coordinate of the target point, (-) , ) In order to map the pixel coordinates, For the preset resolution parameter(s), And As the start coordinates corresponding to the perception range parameters, The height of the map is configured for the region of interest, Is a round down function.
7. The method of claim 1, wherein in the region of interest map, the pixel attribute values are color values, wherein pixels of the region of interest are configured as first color values and pixels of the non-region of interest are configured as second color values; the querying the pixel attribute value corresponding to the mapped pixel coordinate in the region of interest configuration diagram comprises: acquiring a color value at the mapped pixel coordinates; and when the color value is the second color value, judging that the pixel attribute value corresponds to a non-interested region, and otherwise, judging that the pixel attribute value corresponds to a interested region.
8. An electronic device, the electronic device comprising: One or more processors, and A memory storing computer program instructions that, when executed, cause the processor to perform the steps of the method of any one of claims 1 to 7.
9. A computer readable medium having stored thereon a computer program/instruction, which when executed by a processor, implements the steps of the method according to any of claims 1 to 7.
10. A computer program product comprising computer programs/instructions which, when executed by a processor, implement the steps of the method of any of claims 1 to 7.

Description

Method, equipment, medium and product for extracting point cloud of interest area based on roadside laser radar Technical Field The application relates to the field of intelligent traffic, in particular to a method, equipment, medium and product for extracting point clouds of a region of interest based on a road side laser radar. Background With the rapid development of Intelligent Transportation Systems (ITS) and vehicle-road cooperative technology (V2X), road-side perception systems play an increasingly important role in traffic management and automatic driving assistance. In a road side sensing unit (RSU), a laser radar is a core sensor for acquiring three-dimensional information of a traffic scene due to the advantages of high precision, strong anti-interference capability and the like. The lidar is capable of scanning in real time dense point cloud data covering the entire field of view, including environmental information of roads, vehicles, pedestrians, and surrounding buildings, trees, etc. However, in practical road side sensing application, since the laser radar has a wide scanning range and a huge data volume, if all point cloud data in the field of view are subjected to subsequent target detection, tracking and classification processing, huge computing resources and transmission bandwidth are consumed, so that the system delay is increased, and the real-time requirement is difficult to meet. In fact, roadside awareness systems typically only need to be concerned with road surfaces and traffic participants above them, so-called "regions of interest" (ROIs), while a large off-road background (e.g. buildings, flower beds, open space) belongs to invalid information. In the prior art, in order to eliminate these ineffective backgrounds, a method of manually calibrating or drawing a region of interest is generally adopted, for example, manually framing and selecting a road range on a point cloud map. The method is time-consuming and labor-consuming, relies heavily on manual experience, and the generated area is usually static and regular in geometric shape, so that the generated area is difficult to accurately fit with the actual complex road trend (such as curves and irregular intersections). Disclosure of Invention The application aims to provide a method, equipment, medium and product for extracting a point cloud of an area of interest based on a road side laser radar, which are at least used for solving the technical problems that in the prior art, the data processing capacity of a road side sensing system is large, the construction of the area of interest depends on manual calibration and the efficiency is low. To achieve the above object, some embodiments of the present application provide the following aspects: In a first aspect, some embodiments of the present application provide a method for extracting a point cloud of a region of interest based on a roadside laser radar, the method comprising: Acquiring three-dimensional point cloud data acquired by a roadside laser radar; traversing each target point in the three-dimensional point cloud data, and calculating the mapping pixel coordinate of the target point in a pre-constructed region-of-interest configuration diagram according to a preset resolution parameter and a preset perception range parameter, wherein the region-of-interest configuration diagram is a two-dimensional aerial view diagram, and different pixel attribute values represent a region of interest and a non-region of interest; And if the pixel attribute value corresponds to the region of interest, reserving the target point as the point cloud of the region of interest. In a second aspect, some embodiments of the application also provide an electronic device comprising one or more processors and a memory storing computer program instructions that, when executed, cause the processors to perform the steps of the method as described above. In a third aspect, some embodiments of the application also provide a computer readable medium having stored thereon computer program instructions executable by a processor to implement a method as described above. In a fourth aspect, some embodiments of the application also provide a computer program product comprising a computer program/instruction which, when executed by a processor, implements the steps of the method as described above. Compared with the related art, in the scheme provided by the embodiment of the application, the rapid filtering of the laser radar point cloud data at the road side is realized by constructing the region of interest configuration diagram offline and combining the online analysis mode. By mapping the three-dimensional point cloud to the query attribute in the preconfigured two-dimensional pixel coordinates, a large number of non-road background point clouds can be rapidly removed with extremely small calculated amount, so that the data amount required to be processed by a subsequent algorithm is remarkably reduc