Camouflaged target detection based on infrared saliency map and visible image fusion

Abstract

In order to effectively improve the detection ability of camouflaged targets and fully utilize the potential of target detection algorithms, this paper proposes a camouflaged target detection method based on the fusion of thermal infrared saliency map and visible image. Firstly, the saliency map of the target’s infrared band is obtained through the instance segmentation network dedicated to the camouflaged target, and then the saliency map is fused with the target’s visible image with pixel-level weighting, thus obtaining the fused image of the same target in two bands, and this fusion method is aimed at the enhancement of specific regions of the camouflaged target. The target detection by YOLOv8, the experimental results show that the method of this paper has a significant improvement in image detection performance, can effectively improve the recognition ability of camouflage targets.