Sight enhancement through video fusion in a surveillance system

In this paper we consider the problem of fusing two video streams acquired by an RGB camera and a sensor operating in the long wave infrared (LWIR). The application of interest is area surveillance and the fusion process aims at enhancing the human perception of the monitored scene. We propose a fusion procedure where the background and the moving objects are separated and fused by means of different strategies. With respect to standard video fusion techniques this approach has the advantage of reducing the computational load and mitigating the rapid brightness variations in the fused video. It is also less sensitive to the presence of noise. We discuss the experimental results obtained on a typical area surveillance scenario and demonstrate the effectiveness of the proposed method. For this purpose, the analysis is carried out both subjectively, in terms of visual quality of the fused video stream and objectively, in terms of standard image quality indexes. The computational load is also evaluated.