Research On The Arithmetic Of Infrared Image Pre-processing

With the development of Infrared Focal Plan Array (IFPA), the infrared imaging technique is applied in the domain of military, industry and commerce and it becomes a technology which is developed in emphasis in many developed countries in the world. But some problems are necessary to solve in the infrared imaging system. As the effect of production material and manufacture technics, infrared detector will bring in lots of noise in the detection of infrared image. Generally, infrared image has some characteristics of low contrast and narrow grayness scope because of the imaging reason of infrared radiation. These problems affect the quality of infrared imaging and make the target not easy to be identified clearly. It is impossible to solve these problems just through improving apparatus. So researching the pre-processing arithmetic of infrared imaging is significant for improving the quality of infrared image.The most important factor among the multitudinous factors of affecting the quality of infrared image is caused by non-uniformity and blind points, which belong to Original Spatial Noise and can't be eliminated by general method of processing noise. The low contrast is another large defect of infrared image. The arithmetic of improving contrast base on space is easy to carry out by hardware. The main works of this paper is provided as follows:1)The imaging procedure of infrared image is analyzed and the characteristics are summarized from the principle of infrared imaging.2) The mechanism of composing and generation of image noise, and the response model of detector are studied. The generating reason of non-uniformity and bad pixels are analyzed , and the measurement of non-uniformity and judgment standard of bad pixels are presented.3)Many kinds of non-uniformity correction arithmetic are studied, the principles of these arithmetic are analyzed. Their applicability and characteristics are summarized and their advantage and disadvantages of these arithmetic are compared. Some of them are simulated by Matlab software.4)Solving bad pixels are key to confirm blind points. Some arithmetic of detecting bad pixels in this paper are introduced .Some problems and improving direction of different arithmetic are analyzed. 5)The arithmetic of histogram equalization, which is used to improve the contrast of image, is studied according to the characteristic of low contrast of infrared image. A kind of equalization arithmetic based on threshold value of plateau histogram is proposed, this arithmetic can improve the contrast of image better than the statistics histogram equalization, the effect of this arithmetic is proposed.