A robust identification and positioning optimization method for guide lights in AUV docking process

Abstract

The present invention discloses a robust identification and positioning optimization method for a guide light in an AUV docking process, comprising: (1) calibrating internal and external parameters of an underwater camera; (2) preprocessing the original image acquired during the AUV docking process; (3) realizing target field of view area cutting according to the camera's internal parameter matrix and the real-time attitude data provided by the attitude sensor; (4) for the image after the target field of view is cut, firstly performing global binarization processing, then performing local binarization processing, determining and retaining the candidate guide light connected domain; (5) performing inverse perspective homography matrix solution in combination with the attitude data of the attitude sensor, and correcting the morphological contour of the guide light to a shape close to the actual shape; (6) solving the centroid according to the candidate guide light connected domain in the image, and further obtaining the centroid position of the guide light. The present invention improves the accuracy of solving the guide light position, comprehensively realizes the accurate identification and positioning of the guide light in the AUV underwater docking process, and has strong robustness and practicality.

Images