Study Of Infrared Image Enhancement Technology

Infrared image enhancement technology is an important technique in ART(autimatic target recongnition) and Infrared Image Target Tracking System. Infrared image is characterized by low-contrast, blur and high-noise. These features will make a very bad impact on subsequent processing, for example, target recognition and tracking. Therefore, the infrared image enhancement processing is necessary. The author, in this paper, attempts to study infrared image enhancement of spatial domain through the perspective of studying infrared image contrast enhancement, noise reduction and infrared image edge enhancement. This paper will be focus on the follows:The producing principle and features of infrared image are studied. In order to reduce the amount of computation of histogram and satisfy the need of real-time system, a kind of fast method is proposed, i.e., interlace row/column histogram statistics. Contrast enhancement of infrared image is based on gray-scale transformation. The paper analyzed and compared the advantages and disadvantages of various commonly used transformation method, and proposed the project against the requirements of Qingdao city public safety system of infrared target recognition algorithm for adaptive histogram equalization.After types, characters and models of noise are analyzed, the basic theory of image filtering enhancement on spatial domain is studied. On the basis of the above, common infrared image smoothing algorithms are studied and compared by experiment. Then, a kind of improved algorithms of gradient Inverse Weighting Smoothing and median filtering algorithms are presented.In order to enhance the edge information of infrared images, gradient algorithm and laplacian algorithm are analyzed and studied, and two kinds of sharpening operator proposed improvements.The paper also gives the Qingdao city pubic safety vehicle road traffic flow statistics,pattern recognition and application cases.