Reseach On Detail Enhancement Algorithm And Implementation For Infrared Images

As a product of infrared technology, infrared images have been used in the area of military, commercial, research and people’s daily lives more and more widely. Normally, the infrared images should be compressed and enhanced without losing the important detail information because of the disadvantages of low contrast, blurred details and high dynamic range. As a result, much attention has been paid to the digital detail enhancement(DDE) technology for the infrared images. These studies of implementation of DDE technology and the related algorithms have significance both in theory and practice.In this paper, the main characteristics of the infrared image, the concept of image digitization, and the gray-scale histogram are introduced. Based on the classical method, especially the DDE technology, a new infrared image detail enhancement algorithm based on the image layering is proposed. The principle of the bilateral filter and image layering are described. According to the different gray distribution of the two parts, different processing methods are chosen. For the basic part, the improved gray linear mapping method is selected; for the detail part, the S-curve gray transform method is proposed, which is beneficial to the detail preservation and enhancement. Finally, the two parts of the image are combined to be exported.Based on the software imaging testing system and PC general processor platform, the effectiveness and the real-time performance of the proposed algorithm are simulated and implemented using C++ and MATLAB program. The indoor and outdoor infrared images are selected as the processing objects. And the gray linear mapping algorithm is chosen for the experiment. According to the subjective and objective evaluation standard of the image quality, the comparisons of the performance of the different objects were discussed. The experimental results show that the proposed algorithm can effectively enhance the detail information. It is concluded that the average gradient and PSNR are 8.915 and 0.0072, respectively, which are much better than the experimental results. Therefore, the superiority of the proposed algorithm was verified.