Vehicle-Localized Camera Calibration for Non-Rigid Body Motion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Self-driving vehicles experience sensing errors due to non-rigid structures causing relative displacements and shaking of sensors, leading to inaccurate detection of surrounding objects and potential control command malfunctions.
Innovation Solution
A camera calibration method combining vehicle chassis localization with multi-sensor fusion, visual-based vehicle body localization, and a dynamic mass-spring-damper model to iteratively determine six-degree-of-freedom positions and calibrate extrinsic camera parameters using high-definition map information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If the vehicle body is designed as a non-rigid structure with suspension system, then the vehicle can adapt to different road conditions and provide comfortable riding, but the sensors mounted on the vehicle body will experience relative displacement and shaking, leading to sensing errors
Solution Approach 1:
The patent introduces an intermediary coordinate system (vehicle body coordinate system) that moves with the vehicle body, and establishes transformation relationships between this coordinate system and the chassis coordinate system. This intermediary framework allows the system to accommodate vehicle body movements while maintaining accurate sensor positioning through coordinate transformations, thus resolving the conflict between adaptability and sensing accuracy.
Solution Approach 2:
The patent dynamically adjusts the extrinsic parameters of cameras by continuously tracking and compensating for vehicle body movements. Instead of using fixed static parameters, the system实时更新 camera positions and orientations based on actual vehicle body displacement and rotation, ensuring accurate sensing despite the non-rigid structure's movements.
2Loss of information
If cameras are mounted on the vehicle body to detect surrounding objects, then the system can obtain visual information for localization and object detection, but the extrinsic parameters of the cameras will vary due to vehicle body swaying, leading to errors in distance estimation
Solution Approach 1:
The patent implements a feedback mechanism where the system continuously monitors vehicle body movements through sensors (accelerometers, gyroscopes, GPS) and uses this information to dynamically correct camera extrinsic parameters. The corrected parameters are then used for accurate object detection and distance estimation, creating a closed-loop system that maintains precision despite vehicle body variations.
Solution Approach 2:
The patent changes the parameters used for camera calibration from static factory-set values to dynamic parameters that are continuously updated based on actual vehicle body position and orientation. By adjusting extrinsic parameters in real-time according to measured vehicle body displacement and rotation, the system maintains accurate distance estimation even as the vehicle body sways on non-rigid suspension.
3Adaptability or versatility
If the vehicle carries different loads causing body height variations, then the vehicle can accommodate cargo, but the relative position of sensors mounted on the vehicle body will change, resulting in detection errors
Solution Approach 1:
The patent treats sensor positions as dynamic variables rather than fixed manufacturing parameters. By continuously measuring actual sensor positions through onboard sensors and updating extrinsic parameters accordingly, the system accommodates load-induced height variations without sacrificing detection accuracy. The manufacturing precision requirement is replaced with real-time position measurement and compensation.
4Adaptability or versatility
If uneven road surfaces cause vehicle body swaying, then the vehicle can traverse various terrains, but the sensors will offset from their calibrated positions, leading to errors in motion state detection
Solution Approach 1:
The patent introduces an intermediary coordinate system that moves with the vehicle body, serving as a mediator between the chassis coordinate system and sensor measurement frames. This intermediary framework allows the system to traverse uneven terrains while maintaining accurate motion state detection through continuous coordinate transformations that account for body swaying.
Solution Approach 2:
The patent uses feedback from inertial sensors (accelerometers, gyroscopes) to continuously monitor vehicle body movements on uneven terrains and dynamically adjust camera extrinsic parameters. This feedback loop ensures that motion state detection remains accurate despite the vehicle body's adaptive swaying response to terrain variations.
Data Source
AI summary
The present disclosure provides a camera calibration method based on vehicle localization, mainly considering that a vehicle is not of an ideal rigid structure, and shaking of a vehicle body during operation, uneven road surfaces, or different loads will cause variations in extrinsic parameters of a camera system, resulting in perceptual information errors. Therefore, through perceptual information from a fused image, an inertial measurement unit, a speedometer, and the like, combined with the road and marking information provided by a high-definition map, the relative motion between the vehicle body and a vehicle chassis is simultaneously described with a mass-spring-damper model, so as to complete the determination of six-degree-of-freedom positions of the vehicle body and the vehicle chassis, and finally the extrinsic parameters of cameras are immediately calibrated with the vehicle chassis as reference coordinates.