Sea moving target detection method based on mixed multi-Gaussian difference model

Abstract

Compared with land and air target detection, the background of video image sequence of sea moving target is more complex and changeable. In addition to considering illumination variation, shoreline targets occlusion, digital swing and other problems, interference factors such as irregular wave fluctuation and a large number of near-shore breaking waves need to be considered. Because of the size, shape, moving speed and the color correlation between the sea background, it is very difficult to detect moving targets of the sea surface. This paper analyzes the sea moving target detection method based on the mixed multi-Gaussian difference model and applies it to the offshore sea-surface video surveillance. By analyzing the characteristics of near-shore sea-surface target image, the complex scene of ocean dynamic change, the characteristics of inter-frame difference and background difference, the real-time updating strategy of mixed multi-Gaussian background models is studied on the basis of traditional frame difference and mixed Gaussian difference target detection methods, and a mixed multi-Gaussian difference detection model is designed. Through theoretical simulation, the traditional interframe difference method, mixed Gaussian difference method and the proposed method is compared and verified. The simulation results show that the improved algorithm model can detect the outline and detail features of moving targets at sea, especially for fast target detection. In addition, the algorithm does not need the prior knowledge of the scene and can update the background model adaptive in real time, has good environmental robustness.