Multispectral Stereo Camera Self-Calibration via Track Feature Registration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The positional relationship between infrared and visible light cameras is unstable due to temperature, humidity, and vibration changes, making joint self-calibration challenging, especially when fewer effective point pairs are obtained through direct feature point matching.
Innovation Solution
A multispectral stereo camera self-calibration algorithm based on track feature registration, which involves capturing continuous frames of moving objects, extracting and matching motion tracks, optimizing track corresponding points, and using these features to correct the calibration results by de-distorting and binocularly correcting images, and superimposing the corrected positional relationship.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If direct feature point matching is used between infrared and visible light cameras, then the calibration process is simple, but fewer effective point pairs are obtained due to different imaging characteristics
Solution Approach 1:
The patent introduces motion tracks as an intermediary element to bridge the gap between infrared and visible light camera imaging characteristics. Instead of directly matching feature points between the two modalities, the system extracts motion tracks from both cameras and uses track-to-track matching as an intermediate step, which is more robust to the differences in imaging characteristics and yields more effective corresponding points
Solution Approach 2:
The patent transitions from 2D feature point matching to 3D motion track matching by utilizing temporal information. By extracting trajectories of moving objects across multiple frames and matching these 3D motion paths, the system overcomes the limitations of 2D feature matching and obtains more reliable corresponding point pairs
2Ease of manufacture
If joint calibration is performed to obtain positional relationship parameters, then the initial calibration is achieved, but the parameters become unstable when temperature and humidity change
Solution Approach 1:
The patent implements dynamic self-calibration by continuously tracking moving objects and updating the positional relationship parameters in real-time. Instead of relying on static calibration parameters that degrade with environmental changes, the system dynamically adjusts the parameters based on current observations of motion tracks, making the calibration adaptive to temperature and humidity variations
Solution Approach 2:
The system establishes a feedback loop where the accuracy of track matching is continuously evaluated and used to refine the calibration parameters. By monitoring the consistency of matched tracks between infrared and visible light cameras, the system detects drift in positional relationships and corrects it through iterative optimization, ensuring long-term parameter stability
3Productivity
If camera collision occurs accidentally, then the operational continuity is maintained, but the positional relationship between camera lenses changes
Solution Approach 1:
The patent implements preliminary detection mechanisms by continuously monitoring the consistency of motion tracks between the two cameras. When a collision occurs, the system can detect the resulting track mismatches and trigger a self-calibration routine, preventing prolonged operation with inaccurate parameters and maintaining overall system reliability
Data Source
AI summary
The present invention discloses a multispectral stereo camera self-calibration algorithm based on track feature registration, and belongs to the field of image processing and computer vision. Optimal matching points are obtained by extracting and matching motion tracks of objects, and external parameters are corrected accordingly. Compared with an ordinary method, the present invention uses the tracks of moving objects as the features required for self-calibration. The advantage of using the tracks is good cross-modal robustness. In addition, direct matching of the tracks also saves the steps of extraction and matching the feature points, thereby achieving the advantages of simple operation and accurate results.