Research On Infrared Image Denoising And Segmentation Algorithm

The infrared imaging technology belongs to the technology of passive detection, which is applied widely to military and civil affairs. Infrared imaging systems have some outstanding characteristics, such as high resolving power, portable configuration, finer transmission specialty and so on, and could obtain abundant information about intuitionistic configuration of objects.Consequently, infrared imaging has become a research focus in recent years. Compared with visible images, most of infrared images have some disadvantages, such as low contrast, blurred edge, high noise and so on because of the limitation of infrared monitor and the affection of the environment. As the infrared imaging technology being used more and more extensively, the requirement with quality of infrared image has enhanced. At present, the method of infrared image denoising mostly focus on space region or frequency region. Considering the special features of the wavelet transform, for example, it's ability of decompose the two-dimensional signals to different scales, it has been widely applied in the field of denoising.A systematic and deep research is conducted in two aspects based on infrared image processing. Firstly, a transverse comparison of infrared images denoising is made, from common filtering methods including space region and frequency region to wavelet threshold. With the wavelet shrinkage threshold denoising theory proposed by Donoho, an improved method is presented, which deal with each coefficient using mean filter and exponential attenuation function in every scale. The new method overcomes the shortcomings of hard threshold function which is discontinuous and soft threshold function which has warp in processing. Secondly, threshold segmentation method is used by infrared image segmentation because of high real-time needing of infrared imaging system. In this paper, the Otsu algorithm is optimized by the penalty function. The simulation results show the new method in the superiority of running speed.