Radar-Camera Point Cloud Alignment Using Optical Axis Calibration

The patent introduces a calibration board as an intermediary object that serves as a common reference for both the millimeter wave radar sensor and the camera. The calibration board includes multiple markers with known positions and reflection characteristics, enabling the system to establish corresponding relationship between point cloud coordinates and image coordinates. This intermediary facilitates accurate alignment between the optical axes of the sensor and camera without requiring direct mechanical coupling.

The patent replaces mechanical alignment methods with electromagnetic wave-based calibration. Instead of physically adjusting the sensor and camera to align their optical axes mechanically, the system uses millimeter wave reflection and image processing to detect marker positions and calculate transformation matrices. This substitution enables software-based calibration that is more precise and easier to implement than mechanical adjustment.

The calibration system is self-calibrating in the sense that it uses the calibration board with known geometry to automatically determine the transformation parameters between coordinate systems. The system performs self-check by detecting multiple markers and computing consistency of transformation matrices. This self-service approach reduces the need for external calibration equipment and manual intervention, simplifying the overall system despite the added calibration capability.

The calibration board is pre-configured with markers at known positions and geometric relationships before the calibration process. This preliminary preparation of reference information allows the calibration algorithm to work with predetermined data, reducing computational complexity during actual calibration execution. The pre-established geometric relationships serve as prior knowledge that guides the alignment process.