Algorithm Of High Dynamic Range Infrared Image Compression Detail Enhancement

Infrared imaging technology has a wide range of applications in the field of military and civilian detection. Because of the high accuracy and sensitivity of high-bit-wide infrared detectors, high dynamic range infrared image has become a new important direction for the development of infrared images. It’s necessary to compress the data to8bits to meet the normal display of the display device and subsequent rapid processing, however, the improper compression may cause the original information lost, namely the detail loss of the display image. Therefore, we must take appropriate compression method to retain the details and to exhibit detail enhancement. In view of this, this thesis focuses on detail enhancement algorithm of the high dynamic range infrared image compression. The main research works are as follows:First, the paper reviews the development status of infrared image detail enhancement and dynamic range compression algorithm, introduces and simulates the classic infrared image compression and detail enhancement algorithms, analyzes the problems of classic algorithms. On this basis, focusing on two detail enhancement algorithms based on unsharp mask ideas, improve the image details visibility and the contrast, increase the peak signal to noise ratio. The two algorithms use different layer processing and compression algorithms, one is mainly about bilateral filtering and platform histogram equalization, the other is retinex and high frequency enhancement. Meanwhile, this paper refrains artifacts by bilateral filtering and retinex, respectively by means of auxiliary adaptive Gaussian filter to eliminate the gradient reversal phenomenon, combining global and local components to eliminate halo phenomenon. In addition, the noise has been suppressed by adaptive templates enhancement, so that achieving better detail enhancement. According to the subjective and objective results, the two algorithms have their own advantages, so they can be selected for different need.Based on the algorithm research, an algorithm verification software platform has been designed. The software contains a variety of detail enhancement algorithms, and several objective image detail enhancement evaluation indicators, so that the merits and applicability analysis of various algorithms become easy and intelligent.