Ultraviolet detection technology, as immediate area of research focus, has been adopted in the fields of fingerprint identification, corona detection and exhaust plume detection. Low-light CMOS, which can work in even 10<sup>-3</sup>lux, is used in visible light channel. The prominent advantage of the dual-channel Ultraviolet/Low-light CMOS camera is the fusion of UV and wide dynamic range visible light information, which can enrich image details and help observers locate the UV targets in the complicated background around the clock rapidly. The paper studied on the component structure of UV ICMOS, imaging driving, the Ultraviolet/Low-light images fusion algorithm and the photon counting algorithm. The one-inch and wide dynamic range CMOS chip with the coupling optical fiber panel are coupled to the UV image intensifier. In consideration of the ultraviolet detection demand, the driving circuit of the CMOS chips is designed and the corresponding program based on Verilog language is written. After analysis and comparison of the characteristics of UV image and Low-light CMOS image, the improved Laplace pyramid fusion algorithm is applied. UV image and Low-light CMOS image are multiscale decompose, and the features in different frequency layer are chosen from either UV image or Low-light CMOS image. The connected components labeling way is utilized for the UV detection and imaging. At last, the detection experiments of the ultraviolet signal are carried out, and the results are given and analyzed.