基于粗糙集的双波段图像自然感彩色融合

o solve the problem that the contrast of hot targets may decrease after conventional natural sense color fusion based on color transfer processing, a natural sense color fusion method based on rough set theory for dual-band images is put forward. According to the grayscale information of the infrared image, we define the gradient, the grayscale and the noise condition attributes, and divide the color transfer resulting image into three pairs of sub-images. Then we distinguish the sub-images by their attribute differences, and enhance them respectively using power transformation method. By comparing the enhancement effects of rough-set-based processing and the processing which linearly or nonlinearly adjusts the red-green opponent color channel of conventional color transfer resulting image, we found that the former method can avoid such problems as non-smooth color transition, resolution reduction of hot target regions, and unobvious enhancement of some hot regions. The results showed that the proposed rough-set-based method can provide color fusion images with both high target detection capability and good overall natural color feeling, which can effectively help viewers to discover hot targets better while keeping good scene perception.