Abstract: Aiming at the problem that the solar-blind ultraviolet image and visible light image based on single photon counting have few similar features, image superposition and image filtering are used to preprocess the solar-blind ultraviolet image. Then the centroid extraction and Hough transform are used to obtain the coordinates of calibration points in both the solar-blind ultraviolet image and the visible light image. The Zhang’s camera calibration method was then employed to calibrate the solar-blind ultraviolet camera and the visible light camera. Finally, a registration method based on the binocular vision imaging principle was utilized to derive the coordinate mapping relationship between the solar-blind ultraviolet image and the visible light image. An image fusion process was performed using a weighted averaging fusion algorithm. A series of imaging and data processing experiments were carried out on the ultraviolet light source, and the experimental results verified the feasibility of the calibration fusion algorithm designed for the solar-blind ultraviolet-visible dual-spectrum imaging detection.

Key words: single photon counting, solar-blind ultraviolet, camera calibration, image registration, image fusion