The PSF Estimation And Image Restoration Based On Image Motion Analysis Of TDICCD

TDICCD is a new type of image sensor which has developed in recent years. In order to achieve light and miniaturization of satellite cameras, a new generation of high-resolution camera has used a push-broom imaging linear TDICCD as the receiving device.Due to the special working of the camera, TDICCD will generate image motion by the Earth's rotation, carrier vibration, the flying speed and so on in astronautics and airborne remote sensing, resulting in image aliasing, lower resolution, edge blur, contrast desalination and so on. If having the relative movement between the camera and scenes, it will lead to the images blurred, Motion Blur, it is a common problem in the imaging process. The pictures come from the plane or the spacecraft, the high-speed object and the flying missiles in the battlefield with the camera, they may exist this phenomenon. Therefore, the image restoration is the important problems which must be solved in the application of aerospace TDICCD camera.Based on this situation, which has caused the image distortion by the uniform rectilinear motion, this paper carefully has analyzed and researched the reason of this problem, puted forward own solutions. Firstly, this paper introduced the working principle and characteristics of TDICCD, On this basis, analyzed of its various performance parameters;Has analyzed when the TDICCD line scan speed and the movement likely not synchronization through image MTF to the image formation quality influence, leaded to the output of images blurred, Motion Blur. Secondly, according to the imaging principle of the motion-blurred image and the mechanism of TDICCD image motion production, established image degradation model and the recovery model, analyzed the space domain and frequency domain expression of the point spread function and the blurred image spectrum diagram, we can know that accurately estimating the point spread function is the key factors of TDICCD blurred image restoration; Through the analysis of blurred images frequency spectrum, has proposed a new point spread function estimation method, carried on the restoration to the motion blurred image by the frequency domain iteration algorithm. Finally, this paper proposed a new algorithm for the fuzzy image restoration basing on Wiener filter and Constrained least squares algorithm, to solve the bulrred image problem of the uniform rectilinear motion in TDICCD imaging process.The topic work confirmed the above algorithm feasibility from the theory and the experiment, but must achieve the practical level also must solve some problems, like enhances the algorithm the operating speed, enable the algorithm to have timely and so on high.