An algorithm of non-continuous gray-scale histogram enhancement based on the visual characteristics

摘要

Infrared thermal image, with many noises, blurred details and large ranges, needs to be processed on human observation applications. As the dynamic range of the infrared detector data is usually much larger than that of the display device format, infrared image needs to be gray-scale compression before displaying to human eyes. In order to achieve a good vision-observation effect and to retain the major information at the same time, it is proposed an algorithm of non-continuous gray-scale histogram enhancement, based on the human visual characteristics. Firstly, through researching on gray-scale characteristics curve of the human visual on the image resolution, it configures a "visual resolution histogram" (VRH), which has non-continuous gray scales determined by the gray-scale characteristics curve. Secondly, it integrates both equalization method and order-mapping themes, to transform infrared thermal image into the format of continuous and even distribution on gray scales in visual resolution histogram. As well as, it proposes a "Central Segment Histogram Enhancement" (CSHE) to keep mean brightness effects on the visual system. Finally, experiments show that the proposed algorithm provides rich layers and good resolutions of displaying image on human view, as well as reduces the adverse phenomenon of scale moderation in conventional histogram equalization theme.