Research on Color Feature Extraction Method of Digital Camouflage Based on Visual Perception

Color is one of its major features of image recognition, also a key factor in affecting camouflage effect. At present, the methods of color feature extraction used take no account of human visual perception, which may result in inaccurate color feature extraction and lead to bad effect of the digital camouflage in the practical application. In this paper, a new method of color feature extraction of digital camouflage based on color histograms combined with human visual perception is proposed. By introducing visual attention function into the traditional color histogram, several dominant colors are obtained as the color features of digital camouflage. The experiment results shows that the color features extracted by our method are more closed to the background images, in order to reduce the detection probability of target.