Study on perceptual evaluation of fused image quality for color night vision

With the development and the application of visible-infrared image fusion techniques in color night vision, evaluating the performance of image-fused algorithms is becoming an important aspect. In this study, three basic visual evaluation factors including target detection, details and colorfulness, were presented for evaluating perceptual quality of color fulsed images, and the degree of correlation between the perceptual quality and the three evaluation factors was investigated. Visual evaluation experiment on the color images fulsed visible and infrared images of ten scenes by four fusion algorithms was conducted. The experimental results show that the fulsed images have good performances in target detection, the correlation coefficient between the perceptual quality and details is 0.89, and the correlation coefficient between the perceptual quality and colorfulness is 0.75. It indicates that details are the main factor and the colorfulness plays an important role when targets are beyond the detected level.