A radar-panoramic camera joint calibration method and system based on engineered coordinate transformation and robust homography estimation
By employing engineered coordinate transformation and robust homography estimation methods, the instability and distortion issues in the calibration process of panoramic cameras and radars are resolved, achieving high-precision calibration results applicable to fields such as maritime monitoring, unmanned agricultural machinery, and mobile robots.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- HARBIN ENG UNIV
- Filing Date
- 2026-04-07
- Publication Date
- 2026-07-03
AI Technical Summary
Existing technologies suffer from problems such as calibration instability, discontinuity across the suture zone, polar distortion interference, and inconsistent zero-point heading during the calibration process between panoramic cameras and radars, leading to inaccurate calibration results.
A method based on engineering coordinate transformation and robust homography estimation is adopted. By collecting the pixel coordinates of panoramic images and radar polar coordinates, registration samples are constructed, and engineering coordinate transformation is performed to establish the periodic mapping relationship between radar azimuth and panoramic image pixel coordinates. Combined with RANSAC robust estimation, the heading offset parameters and homography matrix are optimized to achieve high-precision calibration.
It achieves high-precision and robust calibration between panoramic cameras and radar, solves the interference of stitching discontinuity and polar distortion, and ensures calibration consistency and accuracy across the entire field of view, making it suitable for widespread application in practical systems.
Smart Images

Figure CN122336014A_ABST