LiDAR-GNSS Relative Pose Calibration With Vertical Constraints
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current calibration algorithms for high-precision mapping using laser radars and navigation positioning systems face challenges in accurately determining the relative pose between the two systems, especially in the vertical direction, due to the lack of vertical data constraints, which limits their ability to calibrate accurately.
Innovation Solution
A method and apparatus that obtain first point cloud data from a laser radar and first pose data from a navigation positioning system, and use pre-collected data from a laser scanner and positioning device to determine the relative pose by converting data into a common coordinate system and employing algorithms like ICP or GICP for accurate calibration.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional calibration algorithms are used, then the calibration process is simple, but the measurement precision of relative pose especially in vertical direction is insufficient
Solution Approach 1:
The calibration process is divided into distinct phases: pre-calibration using a laser scanner to establish accurate spatial relationships, and on-road calibration using the navigation positioning system. This segmentation allows each phase to specialize in specific calibration aspects, improving overall precision without requiring the entire system to be overly complex simultaneously.
Solution Approach 2:
The laser scanner performs preliminary calibration to pre-determine the spatial relationship between the navigation positioning system and the road surface. This preliminary action provides accurate initial parameters (position, attitude, elevation angle) that constrain the subsequent on-road calibration, thereby improving measurement precision while keeping the real-time calibration system simpler.
2Measurement precision
If pre-collected data from laser scanner is used, then vertical direction calibration accuracy is improved, but data collection time and processing complexity increase
Solution Approach 1:
The laser scanner collects point cloud data and determines spatial relationships in advance during pre-calibration. This preliminary action transfers the time-consuming data collection to a separate phase, allowing the actual on-road calibration to proceed faster using the pre-established reference framework.
Solution Approach 2:
The calibration workflow is segmented into offline pre-calibration (laser scanner data collection and processing) and online calibration (navigation system calibration using pre-determined parameters). This segmentation moves heavy processing to when time is less critical, reducing time loss during actual calibration operations.
3Measurement precision
If ICP or GICP algorithms are employed, then calibration accuracy is enhanced, but computational complexity and processing time increase
Solution Approach 1:
The laser scanner pre-determines the spatial relationship and provides initial calibration parameters before on-road calibration begins. This preliminary action provides excellent initial guesses for ICP/GICP algorithms, reducing the number of iterations needed and thereby reducing computational complexity while maintaining high accuracy.
Solution Approach 2:
The calibration system uses different algorithmic approaches for different calibration phases: the laser scanner provides precise local spatial relationships in pre-calibration, while the navigation system performs local refinements during on-road calibration. This localized application of quality-appropriate methods optimizes the balance between accuracy and complexity.
Data Source
AI summary
Embodiments of the present disclosure disclose a method for calibrating a relative pose, a device, and a medium. The method includes: obtaining first point cloud data of a scene collected by the laser radar in an automatic driving mobile carrier and first pose data collected by the navigation positioning system in the automatic driving mobile carrier; and determining the relative pose between the laser radar and the navigation positioning system based on the first point cloud data, the first pose data, second point cloud data pre-collected by a laser scanner in the scene and second pose data pre-collected by a positioning device in the scene.