Parameter calibration method and apparatus of multi-line laser radar, device and readable medium

Patent metadata

Field	Value
Publication number	US-11714174-B2
Application number	US-201816117953-A
Country	US
Kind code	B2
Filing date	Aug 30, 2018
Priority date	Sep 17, 2017
Publication date	Aug 1, 2023
Grant date	Aug 1, 2023

Abstract

The method comprises: creating a 3D scenario model of a calibration scenario, based on data of the calibration scenario collected by an already-calibrated 3D scanner; obtaining point cloud data collected by a to-be-calibrated multi-line laser radar at a plurality of position points in the calibration scenario respectively; aligning the point cloud data collected by the multi-line laser radar at the respective point points and point cloud data in the 3D scenario model into the same coordinate system on a principle that the respective position points are aligned with corresponding position points in the 3D scenario model; building a target function between data of points collected by the multi-line laser radar and data of matching points nearest to the points in the 3D scenario model, under the same coordinate system; calibrating the parameters of the multi-line laser radar according to the target function.

First claim

What is claimed is:

1. A method of calibration of internal and/or external parameters of a multi-line laser radar in view of different calibration scenario, the external parameters being related to both the multi-line laser radar and a mounting machine mounted with the multi-line laser radar, wherein the method comprises:
creating a 3D scenario model of a calibration scenario, based on data of the calibration scenario collected by an already-calibrated 3D scanner;
obtaining point cloud data collected by a to-be-calibrated multi-line laser radar at a plurality of position points in the calibration scenario respectively, wherein a precision of the already-calibrated 3D scanner is larger than the precision of the to-be-calibrated multi-line laser radar;
aligning absolutely the point cloud data collected by the multi-line laser radar at the respective point points and point cloud data in the 3D scenario model into the same coordinate system on a principle that the respective position points are aligned with corresponding position points in the 3D scenario model;
building an objective function between data of points collected by the multi-line laser radar and data of matching points which are nearest to the respective points collected by the multi-line laser radar in the 3D scenario model under the same coordinate system, wherein the objective function defines a point-to-point correspondence between a point collected by the multi-line laser radar and a corresponding matching point in the 3D scenario model;
calibrating the parameters of the multi-line laser radar according to the objective function.

2. The method according to claim 1 , wherein the aligning the point cloud data collected by the multi-line laser radar at the respective point points and point cloud data in the 3D scenario model into the same coordinate system on a principle that the respective position points are aligned with corresponding position points in the 3D scenario model specifically comprises:
aligning respective position points to corresponding positions in the 3D scenario model;
obtaining a conversion parameter at respective position points after the alignment from a coordinate system of a mounting machine mounted with the multi-line laser radar to a coordinate system of the 3D scenario model;
according to the conversion parameter at respective position points from the coordinate system of the mounting machine mounted with the multi-line laser radar to the coordinate system of the 3D scenario model, and an internal parameter and an external parameter of the multi-line laser radar, converting the point cloud data collected by the multi-line laser radar at respective position points to corresponding coordinates in the 3D scenario model, so that the point cloud data collected at respective position points and the 3D scenario model belong to the same coordinate system.

3. The method according to claim 2 , wherein the step of, according to the conversion parameter at respective position points from the coordinate system of the mounting machine mounted with the multi-line laser radar to the coordinate system of the 3D scenario model, and an internal parameter and an external parameter of the multi-line laser radar, converting the point cloud data collected by the multi-line laser radar at respective position points to corresponding coordinates in the 3D scenario model specifically comprises:
taking r j as raw data at any point j in the point cloud collected by the multi-line laser radar at the i th position point, and calculating coordinates of the point j in the coordinate system of the mounting machine by using the equation p b j =T b,s L(r j ; I), according to the internal parameter I of the multi-line laser radar, the conversion function for converting the raw data into the multi-line laser radar coordinate system, and the conversion parameter from the coordinate system of the multi-line laser radar to the coordinate system of the mounting machine mounted with the multi-line laser radar, wherein r j ∈R i , R i is a set of raw data collected by the multi-line laser radar at the i th position; L(r j ; I) is the conversion function from the raw data to the multi-line laser radar coordinate system; T b,s is a set of conversion parameters from the coordinate system of the multi-line laser radar to the coordinate system of the mounting machine mounted with the multi-line laser radar, T b,s ={T b,s i }, wherein T b,s i is the conversion parameter from the coordinate system of the i th multi-line laser radar to the coordinate system of the mounting machine mounted with the multi-line laser radar; p b j is the coordinates of the point j in the coordinate system of the mounting machine;
calculating coordinates of the point j in the coordinate system of the 3D scenario model by using equation p m j =T m,b i p b j according to a conversion parameter T m,b i at the i th position point from the coordinate system of the mounting machine to the coordinate system of the 3D scenario model, and coordinates p b j of the point j in the coordinate system of the mounting machine; wherein p m j is coordinates of the point j in the coordinate system of the 3D scenario model.

4. The method according to claim 3 , wherein the building an objective function between data of points collected by the multi-line laser radar and data of matching points nearest to the points in the 3D scenario model, under the same coordinate system specifically comprises:
setting a weight w(p)=f(c(p)) and a normal direction n(p) of points p in the 3D scenario model; wherein c(p) is a scalar function, which measures a curve degree of a model surface adjacent to point p; function f( ) is a monotone function to map c(p) to between 0 and 1; a value of w(p) is closer to 1 in a flatter region of the surface adjacent to the point p measured by c(p), otherwise the value of w(p) is closer to 0 in a more curved region of the surface adjacent to the point p measured by c(p); n(p) is the normal direction of the point p;
obtaining coordinates m j of a matching point nearest to coordinates p m j of the point j in the 3D scenario model collected at the i th position point, and obtaining a weight w(m j ) and a normal direction n(m j ) of the matching point m j ;
building an objective function using the following formula according to the coordinates p m j of the point j in the coordinate system of the 3D scenario model, the coordinates m j of the matching point, and the weight w(m j ) and the normal direction n(m j ) of the matching point m j :
E ( T m,b ,T b,s ,I )=Σ i Σ j ρ( w ( m j ) n T ( m j )( m j −p m j )), m j ∈M

where E(T m,b , T b,s , I) is an objective function; ρ(·) is a loss function; n T (m j ) is transposition of the normal direction n(m j ) of the matching point m j ; M is a set of points in the 3D scenario model; T m,b ={T m,b i } and is a set of conversion parameters T m,b i of all i position points from the coordinate system of the mounting machine to the coordinate system of the 3D scenario model.

5. The method according to claim 4 , wherein the calibrating the parameters of the multi-line laser radar according to the objective function specifically comprises:
optimizing a conversion parameter T b,s in the objective function from the coordinate system of the multi-line laser radar to the coordinate system of the mounting machine mounted with the multi-line laser radar, to make the objective function converge;
jointly optimizing the T b,s and internal parameter I in the objective function to enable the objective function to take a minimal value;
obtaining a value of the corresponding T b,s and internal parameter I when the objective function takes the minimal value, respectively as the external parameter and int

Assignees

• Apollo Intelligent Driving Tech Beijing Co Ltd

Inventors

• Wang Bosheng
• Xie Yuanfan
• Wang Jun
• Yan He
• Wang Liang

Classifications

• G01S7/497 Primary

  Means for monitoring or calibrating · CPC title

• G01S17/42

  Simultaneous measurement of distance and other co-ordinates (indirect measurement G01S17/46) · CPC title

• G01S7/40 Primary

  Means for monitoring or calibrating · CPC title