Algorithm for Enhancing Salient Objects in Color Fusion of Night Vision Images Based on Spectral Residual
夜视图像彩色融合中基于谱残差的显著目标增强算法

According to the characteristics of human visual systems, an algorithm for enhancing the salient objects in color fusion of night vision images based on spectral residual is proposed. In the color fusion process, the range of the salient object is determined by the spectral residual and the salient object is located with a window of 5?5 pixels. When the mean value in this window is greater than the global mean value, the pixels are located on the salient object. Then, the salient object is highlighted in red color so as to enhance the color fusion image. The experimental result shows that the new algorithm is stable and practical and is more effective in the enhancement of salient objects.