An Improved Enhancement Method for Infrared Images

Abstract

Infrared images play important roles in the national defense and civil applications. The enhancement of the infrared images aims to improve the visual quality of images so that the result is more suitable for specific applications or an observer. In this paper, we explored the histogram equalization method which is a common technique for image enhancement, and proposed an improved enhancement method for infrared images. More specifically, we proposed a method to automatically select the threshold for histogram segmentation by an iterative method, and process the pixel brightness locally by considering the visual factor. We tested our method on different types of infrared images, and quantitatively compare our method with previous methods. The results show that out method improved visual quality of the images and also provide higher entropy and peak signal noise ratio.