Onboard Camera Calibration Using Vehicle Body Features
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing calibration methods for onboard cameras in vehicles struggle to achieve high accuracy when the road surface is not flat, affecting the precision of distance calculations and camera orientation calibration.
Innovation Solution
A calibration device installed in vehicles that extracts feature points from both the road surface and the vehicle body from captured images, using these points to calibrate camera orientation parameters, including triaxial angles and elevation, even on non-flat road surfaces.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If calibration is performed using road surface features, then calibration can be done during vehicle travel, but accuracy deteriorates when the road surface is not flat
Solution Approach 1:
The calibration process is segmented into two independent feature extraction paths: road surface feature points and vehicle body feature points. This segmentation allows the system to use multiple feature sources, reducing dependence on road surface flatness while maintaining calibration efficiency during vehicle travel.
Solution Approach 2:
Vehicle body feature points serve as an intermediary reference that is independent of road surface conditions. By introducing this intermediary calibration reference, the system can achieve accurate calibration even when road surfaces are uneven, while still performing calibration during vehicle operation.
2Ease of manufacture
If external parameters are calibrated in the factory before shipping, then the calibration process is simplified, but the parameters become inaccurate when the vehicle orientation changes during operation
Solution Approach 1:
The system performs preliminary calibration in the factory to establish baseline external parameters, simplifying the manufacturing process. Additionally, it prepares the capability for in-vehicle calibration by equipping the camera system to capture both road surface and vehicle body features, enabling accuracy correction when vehicle orientation changes during operation.
Solution Approach 2:
The calibration system transitions from a static factory calibration approach to a dynamic in-vehicle calibration approach. The system can adaptively recalibrate external parameters during vehicle operation based on captured images, maintaining parameter accuracy despite changes in vehicle orientation while keeping the overall process simple.
3Measurement precision
If multiple feature points from both road surface and vehicle body are extracted, then calibration accuracy is improved, but the device complexity increases
Solution Approach 1:
The image processing system is designed with multi-functionality to extract both road surface feature points and vehicle body feature points using the same hardware and software platform. This universal approach improves calibration accuracy by utilizing multiple feature sources without requiring separate dedicated systems, thereby limiting the increase in device complexity.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The present invention calibrates parameters indicating an orientation of a camera even if a road surface on which a vehicle travels is not flat. A calibration device 100 is installed in a self-vehicle 1A, and includes a feature point extraction unit 154 configured to extract, from a captured image of a camera 4 installed in the self-vehicle 1A, a feature point included in a region of the captured image in which a road surface around the self-vehicle 1A is captured; a self-vehicle body recognition unit 153 configured to extract, from the captured image, a vehicle body feature point included in a region of the captured image in which a vehicle body 10 of the self-vehicle 1A is captured; and an external parameter estimation unit 158 configured to calibrate, based on the feature point extracted by the feature point extraction unit 154 and the vehicle body feature point extracted by the self-vehicle body recognition unit 153, external parameters 125 indicating a mounting state of the camera 4 relative to the self-vehicle 1A.