Infrared target detection in backlighting maritime environment based on visual attention model

Abstract

The backlighting environment (BE) is a common condition in the process of detecting maritime targets. The reason for the phenomenon of backlighting at maritime is that when the weather is sunny, the target at sea is just between the sunlight and the infrared camera, which causes the radiation energy of sunlight reflected into the infrared camera through the sea surface to be much larger than that reflected into the infrared camera through the target. Therefore, the gray value of the background of the infrared image is larger than the gray value of the target in the BE. In view of this situation, the main content of this paper is to design a robust infrared maritime target detection algorithm based on visual attention model (VAM) by analyzing the characteristics of the target in the BE. First and foremost, a preprocessing method of Gauss difference is proposed to improve the gray value and contrast of the target in strong ocean waves backlighting image. Secondly, a suitable modality will be automatically selected according to the intensity of different ocean waves. Moreover, we adopt the center-surround difference (CSD) that highlight the saliency of the target instead of the CSD at multiple scales in the VAM. Finally, the saliency map is segmented by an adaptive threshold method OTSU, and the integrity of the target area is guaranteed. Thus, the goal of successfully detecting the target is achieved. Experimental results show that the proposed algorithm remarkably improve the accuracy and efficiency of searching maritime targets in the BE.