CN-116777936-B - Line channel self-adaptive segmentation method and system for unmanned airborne laser radar

CN116777936BCN 116777936 BCN116777936 BCN 116777936BCN-116777936-B

Abstract

The invention provides a line channel self-adaptive segmentation method and system of an unmanned airborne laser radar, wherein the self-adaptive segmentation method comprises the steps of resolving point cloud data of the laser radar in real time, extracting to obtain a plurality of power line point clouds, fitting a power line linear equation corresponding to each power line point cloud respectively, calculating to obtain a plurality of power line lines by using the power line linear equation, clustering the mutually parallel power line lines into the same power line group according to a parallel line clustering algorithm, segmenting the same power line group into a plurality of line groups, calculating the minimum line spacing of the power line lines, segmenting the rest power line lines into the line groups according to the minimum line spacing and the channel average line spacing of the line groups, and aggregating all the power line lines of the same line group into the same line channel when all the power line lines are segmented. The technical scheme of the invention can solve the problems that in the prior art, all the line channels are difficult to automatically identify and the line-imitating flight is difficult to automatically simulate according to the intention of a user.

Inventors

ZHANG YONG
LI ZHIBO
GUO YANMING
WU FANGFANG
SUN XILIANG
ZHANG HENG
CHEN LINHAI
GAO SHANG
XU GUANGCAI

Assignees

北京数字绿土科技股份有限公司
深圳绿土智新科技有限公司

Dates

Publication Date: 20260505
Application Date: 20230423

Claims (9)

1. The self-adaptive segmentation method for the line channel of the unmanned airborne laser radar is characterized by comprising the following steps of: in the flight process of the unmanned aerial vehicle, resolving point cloud data acquired by a laser radar in real time, and extracting to obtain a plurality of power line point clouds; Fitting a power line linear equation corresponding to each power line point cloud in the plurality of power line point clouds according to a preset line fitting algorithm; Selecting any power line as a reference line by using the power line linear equation, performing direction correction on other power line lines by using the reference line, and clustering the power line lines which are parallel to each other into the same power line group according to a parallel line clustering algorithm; dividing the same power line group into a plurality of line sets according to the dividing line spacing of the power line lines; For each line set in the line sets, calculating the minimum line spacing between the remaining power line straight lines in the same power line group and the line set, and dividing the remaining power line straight lines into the line sets according to the size relation between the minimum line spacing and the channel average line spacing of the line sets; when all the power lines in the same power line group are linearly segmented, linearly aggregating all the power lines in the same line set to the same line channel; The step of calculating a minimum line spacing between a remaining power line in the same power line group and the line set for each line set in the plurality of line sets, and dividing the remaining power line into the line sets according to a magnitude relation between the minimum line spacing and a channel average line spacing of the line sets, includes: Selecting a power line from the power line group and adding the power line to the current divided line set; Respectively calculating the minimum line distance between each remaining power line straight line and the line set in the same power line group; Respectively judging whether the minimum line spacing between each remaining power line and the line set in the same power line group is smaller than the split line spacing; if the minimum line spacing is smaller than the split line spacing, judging whether the minimum line spacing is smaller than the product of the channel average line spacing and the channel spacing factor of the line set; if the minimum line spacing is less than the product of the average line spacing and the channel spacing factor in the line set, the remaining power line straight line is added to the line set; if the minimum line spacing is greater than or equal to the product of the average line spacing and the channel spacing factor in the line set, reserving the remaining power line straight lines to the same power line group; increasing the separation line spacing, and re-executing the step of judging whether the minimum line spacing between each remaining power line and the line set in the same power line group is smaller than the separation line spacing or not until the separation line spacing is larger than a preset maximum line spacing; And when all the power lines in the same power line group are linearly segmented, aggregating all the power lines in the same line set to the same line channel.
2. The line channel adaptive segmentation method according to claim 1, wherein the step of extracting a plurality of power line point clouds by resolving point cloud data acquired by the lidar in real time includes: acquiring point cloud data of a target power line area by using the laser radar; According to the relation between the laser radar coordinate system and the world coordinate system, calculating the point cloud data to obtain point cloud data under the world coordinate system; And extracting a plurality of power line point clouds from the point cloud data under the world coordinate system according to a preset point cloud characteristic extraction algorithm.
3. The line channel adaptive segmentation method according to claim 1, wherein the step of fitting the power line linear equation corresponding to each of the plurality of power line point clouds according to a predetermined line fitting algorithm, respectively, comprises: Selecting coordinates of any two laser points in each power line point cloud by using a RANSAC algorithm; calculating coordinates of the arbitrary two laser points to obtain a fitting central line of the power line point cloud; According to a straight line distance calculation formula, distances from other laser points in the power line point cloud to the fitting center line are respectively obtained; Counting the points of which the distances between the other laser points and the fitting center line are smaller than or equal to a preset shortest distance value; and re-selecting coordinates of any two laser points to calculate a fitting center line of the power line point cloud, and taking the re-calculated fitting center line as a power line linear equation corresponding to the power line point cloud until the distance between the other laser points and the fitting center line is the largest and the point number which is smaller than or equal to a preset shortest distance value is the largest.
4. The line channel adaptive segmentation method according to claim 1, wherein the step of selecting any one of the power line lines as a reference line using the power line equation, performing direction correction on other power line lines using the reference line, and clustering the power line lines parallel to each other into the same power line group according to a parallel line clustering algorithm comprises: Calculating a power line corresponding to the power line point cloud by using a power line linear equation; selecting any one power line from all power line lines as a reference line of power line grouping, and respectively calculating power line included angles between other power line lines in all power line lines and the reference line; judging whether the included angle of the power lines of the other power lines and the reference line is smaller than or equal to a preset included angle threshold value; If the included angle of the power lines is smaller than or equal to a preset included angle threshold value, determining that the other power line straight lines are parallel to the reference straight line, and adding the other power line straight lines into a power line group where the reference straight line is located; If the included angle of the power lines is larger than a preset included angle threshold, any power line straight line is selected from all power line straight lines with the included angle of the power lines larger than the preset included angle threshold to serve as a reference straight line, and the step of judging whether the included angle of the power lines of other power lines and the reference straight line is smaller than or equal to the preset included angle threshold is repeated until all the power line straight lines which are parallel to each other are clustered to the same power line group.
5. The line channel adaptive segmentation method according to claim 1, wherein after the step of clustering the parallel power lines into the same power line group according to a parallel line clustering algorithm, the method further comprises the step of de-duplicating the repeated or split power line lines in the same power line group according to the distance between the power line lines in the same power line group, wherein the step of de-duplicating the repeated or split power line lines in the same power line group according to the distance between the power line lines in the same power line group comprises the steps of: Selecting any one power line straight line in the same power line group as a reference straight line of a container, and calculating the distance between all other power line straight lines in the same power line group and the reference straight line; respectively judging whether the distance between each other power line straight line and the reference straight line is smaller than or equal to a preset straight line distance; If the distance is smaller than or equal to a preset linear distance, determining that the other power line straight lines are power line straight lines which are repeated or split with the reference straight line, and adding the other power line straight lines into a container corresponding to the reference straight line; If the distance is greater than the preset linear distance, selecting any power line straight line from all power line straight lines with the distance greater than the preset linear distance as a reference straight line of a container, and repeatedly judging whether the distance between the other power line straight lines and the reference straight line is smaller than or equal to the preset linear distance or not until all the power line straight lines in the same power line group are added into the container corresponding to the reference straight line, wherein each reference straight line corresponds to a unique container; and traversing the power line straight lines in all the containers, and respectively selecting the power line straight lines with the highest height from each container for merging to obtain the same power line group after the duplication removal.
6. The line channel adaptive segmentation method according to claim 1, wherein after the step of linearly clustering the mutually parallel power lines into the same power line group according to a parallel line clustering algorithm, the method further comprises: Judging whether a plurality of circuit channels exist in the same power line group according to the dividing line spacing of the power line straight line; If a plurality of line channels exist in the same power line group, executing a step of dividing the same power line group into a plurality of line sets according to the dividing line spacing of the power line straight line; And if a plurality of circuit channels do not exist in the same power line group, linearly aggregating all power lines in the same power line group into the same circuit channel.
7. The line channel adaptive segmentation method according to claim 6, wherein the step of calculating a minimum line spacing between each remaining power line and the line set in the same power line group, respectively, comprises: taking a power line straight line selected from the power line groups as an initial straight line of the line set; calculating the distance between each remaining power line in the same power line group and the initial line of the line set as the minimum line spacing, or Calculating distances between each remaining power line straight line in the same power line group and all straight lines in the line set respectively; selecting the minimum distance between each remaining power line straight line and the distance between each remaining power line straight line and all straight lines in the line set as the minimum line distance; When the separation line spacing is larger than a preset maximum line spacing, judging whether a residual power line straight line exists in the same power line group or not; and if the remaining power line straight lines exist in the same power line group, merging all the remaining power line straight lines in the same power line group into an independent line channel.
8. A line channel adaptive segmentation system of an unmanned airborne laser radar, characterized in that the line channel adaptive segmentation method of the unmanned airborne laser radar according to any one of claims 1 to 7 is used, comprising: The power line point cloud extraction module is used for resolving point cloud data acquired by the laser radar in real time in the flight process of the unmanned aerial vehicle, and extracting a plurality of power line point clouds; The power line linear equation fitting module is used for respectively fitting power line linear equations corresponding to each power line point cloud in the plurality of power line point clouds according to a preset line fitting algorithm; The power line linear clustering module is used for selecting any power line linear as a reference linear by using the power line linear equation, carrying out direction correction on other power line linear by using the reference linear, and clustering the parallel power line linear into the same power line group according to a parallel line clustering algorithm; The line set dividing module is used for dividing the same power line group into a plurality of line sets according to the dividing line spacing of the power line straight line; A minimum line spacing calculation module, configured to calculate, for each line set in the plurality of line sets, a minimum line spacing between a remaining power line straight line in the same power line group and the line set; The power line straight line segmentation module is used for linearly segmenting the residual power line into the line set according to the magnitude relation between the minimum line spacing and the channel average line spacing of the line set; And the power line linear clustering module is used for converging all power lines of the same line set to the same line channel when all power lines in the same power line group are cut off.
9. A line channel adaptive segmentation system of an unmanned airborne laser radar, comprising a memory, a processor and a line channel adaptive segmentation program of an unmanned airborne laser radar stored on the memory and running on the processor, wherein the tracking channel adaptive segmentation program when executed by the processor implements the steps of the line channel adaptive segmentation method according to any one of claims 1 to 7.

Description

Line channel self-adaptive segmentation method and system for unmanned airborne laser radar Technical Field The invention relates to the technical field of laser radars, in particular to a line channel self-adaptive segmentation method and system of an unmanned airborne laser radar. Background The laser radar is widely used in the field of automatic power line inspection for operation. And acquiring point cloud data of the power line through a laser radar, so as to achieve the aim of inspection. The unmanned aerial vehicle carries the laser radar to automatically patrol the power line, so that the working intensity can be effectively reduced, the intelligent degree is improved, and stable patrol of the line navigation belt is facilitated. In the line-imitating flight process of the unmanned aerial vehicle, the situation that a plurality of line channels exist in a preset distance range below or in front of the unmanned aerial vehicle is unavoidable. In order to select a proper line channel from a plurality of line channels for line-imitating flight, all the line channels need to be accurately identified in advance so as to automatically imitate line-imitating flight according to the intention of a user. In particular, in some scenarios, there are many parallel lines, and even a situation in which the main network and the distribution network are parallel occurs simultaneously, which results in a greater uncertainty of the unmanned aerial vehicle in automatic flight. Eventually, a wrong channel is tracked or a channel area which is deviated from a channel area which is actually required to fly is caused, and operation collection is affected. Disclosure of Invention The invention provides a line channel self-adaptive segmentation scheme of an unmanned aerial vehicle laser radar, and aims to solve the problems that in the prior art, a plurality of line channels exist, and when an unmanned aerial vehicle automatically flies, a large uncertainty exists, so that tracking errors or deviation channel areas are caused, and unmanned aerial vehicle acquisition operation is influenced. In order to solve the above problems, according to a first aspect of the present invention, a line channel adaptive segmentation method of an unmanned airborne laser radar is provided, including: in the flight process of the unmanned aerial vehicle, resolving point cloud data acquired by a laser radar in real time, and extracting to obtain a plurality of power line point clouds; Fitting a power line linear equation corresponding to each power line point cloud in the plurality of power line point clouds according to a preset line fitting algorithm; selecting any power line as a reference line by using a power line linear equation, performing direction correction on other power line lines by using the reference line, and clustering the parallel power line lines into the same power line group according to a parallel line clustering algorithm; dividing the same power line group into a plurality of line sets according to the dividing line spacing of the power line lines; for each line set in the line sets, calculating the minimum line spacing between the rest power line straight lines in the same power line group and the line set, and dividing the rest power line straight lines into the line sets according to the magnitude relation between the minimum line spacing and the channel average line spacing of the line sets; When all the power lines in the same power line group are linearly divided, all the power lines in the same line set are linearly aggregated to the same line channel. Preferably, in the line channel adaptive segmentation method, the step of resolving point cloud data acquired by the laser radar in real time and extracting a plurality of power line point clouds includes: collecting point cloud data of a target power line area by using a laser radar; calculating point cloud data according to the relation between the laser radar coordinate system and the world coordinate system, and obtaining point cloud data under the world coordinate system; and extracting a plurality of power line point clouds from the point cloud data under the world coordinate system according to a preset point cloud characteristic extraction algorithm. Preferably, in the above line channel adaptive segmentation method, the step of fitting the power line linear equation corresponding to each of the plurality of power line point clouds according to a predetermined line fitting algorithm includes: Selecting coordinates of any two laser points in each power line point cloud by using a RANSAC algorithm; Calculating coordinates of any two laser points to obtain a fitting central line of the power line point cloud; according to a linear distance calculation formula, distances from other laser points in the power line point cloud to the fitting center line are respectively obtained; Counting the points of which the distances between other laser points and the fitting