External parameter calibration method of outdoor pole camera based on drone assistance

In order to solve the problem of insufficient accuracy and poor applicability of traditional calibration methods in complex scenes, this paper proposes an extrinsic calibration method based on drone assistance. The target pixel coordinates in the images taken by the camera and the drone camera are extracted through target detection. The drone image and the map image are aligned using spatial alignment technology, and the target pixel coordinates in the drone image are converted into geographic coordinates. Finally, by matching the pixel coordinates and geographic coordinates of the same target, the camera extrinsics are calculated and optimized to achieve high-precision calibration. Experimental results show that the calibration accuracy of this method is better than that of traditional methods in a large field of view, the reprojection error is less than 3 pixels, and the geographic error is controlled within 0.5 meters. This study provides a calibration method that is easy to operate and reproduce, overcomes the limitations of traditional calibration technology in large-scale scenes, and provides a feasible solution for high-precision target positioning in complex scenes.