Research On Infrared Image Non-uniformity Correction And Image Enhancement Technology

Infrared imaging system has the characteristic of strong anti-jamming performance，target recognition capability and also passive work. It has been widely used in military andcivilian fields. The imaging technique has broad prospects for development. So, there isimportant significance for the development of a high performance infrared imaging system.Infrared focal plane array is a key part of thermal imaging system. Its performanceaffects the operation of the system directly. However, limited by current technology, infraredfocal plane arrays has the problem of non-uniformity, and infrared images have the features oflow contrast; it is hard to distinguish between target and background. This paper presentsmany effective methods to solve these problems which exist in the infrared focal plane arrays.In this paper, more contexts are given:(1) The results obtained by traditional neural network correction algorithm, have theproblems of blurring effect and some existing artifacts. After analysis the reasons of theproblems that the traditional neural network correction algorithm may cause, in this article animproved algorithm is proposed: replace the mean filter, which used in the traditionalalgorithm by the nonlinear filter. And the parameters of nonlinear filter are variable accordingto the local image detail changes in the improved algorithm. The corrected images by theimproved algorithm achieve not only a significant improvement in image clarity, but alsoeffectively eliminate the problem of artifacts caused by the traditional algorithms.(2) After analyzed the cause of blind pixels and the difference between blind pixels andpoint targets, an effective detection algorithm was proposed which based on2D linearextrapolation theory. The experiment shows that the proposed detection algorithm was moreeffective than the original algorithm, it verify the effective of the detection algorithm.(3) Infrared images have the features of dark and low contrast; it is hard to distinguishbetween target and background. This article combining with the fuzzy set theory, gray-scaletransformation theory and multi-resolution image fusion theory, proposed a new imageenhancement algorithm. The algorithm not only can improve the contrast of infrared images,but also can keep details of original images. So, the edge information of the processed imagemore apparent. In this article we achieve the purpose of Infrared image non-uniformity correction, Blindpixel compensation and Image Enhancement. And also lay the foundation for our future work.