Research On The Real-time Image Pre-processing Of TDICCD Camera With High Resolution In Space

Optics imaging in space is the hot research topic in the area of optic sensing. Ittakes the most important role in space imaging because of its high resolution, wideview field, and high reliability. CCD (charged coupled device) is the basis of theoptics imaging in space. Recent years, the research on optics imaging is focused onthe optics lens, mechanics structures, the control architectures and the PCB or circuits.It doesn’t give enough emphasis on the image processing. So it is necessary that wereview the image quality of optics imaging system in space in the point of imagingreal-time process in software, and get higher quality and performances images.This paper focuses on the image real-time process and discusses the techniquesof TDICCD image pre-processing. It analyzes the actual imaging problems in spaceprojects, and gives some advice and processing methods. The existing questions thatcan be resolved by digital image processing in the wide field and high resolutionimaging system are the follows: image dis-match in mosaic CCDs, the pixel responsenon-uniformity, the cross-talk between high frequency signals, noisy of system, imageblur by motions, low dynamic range of imaging system and so on.Chapter1introduces the development of imaging techniques in space, and theimage processing techniques in this real-time processing area. Chapter2introduces the architecture of the TDICCD imaging system, and somerelated techniques. It compares different design architectures and their performances.It discusses the factors that impact the imaging performances, and give theillustrations that enhance the imaging ability by image processing.Chapter3depicts the real-time denoise methods in space. It analyzes thecommon ways of dealing with the CCD noise by hardware and software, andcompares the costs and performances of different algorithms. It designs the methodsof wavelet denoise and mid-value denoise by lines. Considering the difficulty betweenthe space region and the frequency region transition, it changes the denoisy method inspace narrow laps on the basis of segment divided, and gives the real-time filterprocess methods. The new method and its hardware architecture adjust the imagingsystem, optimize the system design, and redefine the standard of judging the real-timeprocess system performances.Chapter4designs a few methods of the image real-time enhancements andimage PRNU correction methods according to the imaging characters. It provides theself-adjustable image enhancement algorithms by line segment divided. On the basisof pixel correction, it gets the module of adjustable algorithm at different parametersof gains and clamps. Finally, it compares the different performances of the algorithmsChapter5researches on the implement methods of the algorithms, it gives thedemands of the FPGA resources and performances. It simplify the algorithms andimplements them. Finally, it gives the costs and merits and verifies the errors ofcalculation and simulations.Chapter6discusses the experiments and test results, it compares the theory andactual differences, give the system indexes of imaging process added. Chapter7givesthe conclusions and the illustrations.This paper develops some image process algorithms, and changes them to lineprocess. Finally, it implements the image enhancement, image filter, image correction,improves the image view effect.The main new tasks in the paper are as follows: （1）Design the linear adjustment module of the pixel rectifies parameters to theTDICCD camera at different gains and clamps. Minimize the parameters stored andenhance the ability of the pixel correction.（2）Give the image gray value stretch method based on line transferring that canbe self adjustable and implement it. It improves the image contrast and dynamic range,generates a good image view effect. It enhances the detail description ability, and getsa good resolution of the imaging system.（3）It designs the psy-mid-value filter based on the statics of line transferring.When denoising the pulse noisy image, it keeps the image in better detail.（4）Designs the image enhancement methods and the denoising methods by theline and image characters training. It decreases the image noise level and reduce thesensitivity of the image to the noise, improves the view effect.The space image real-time processing is a very important and challenge area inthe image processing area. When we use the common methods to the real-time andbased-on-architecture of the image data systems, they are more or less limited by theresource or the processing time or the others factors. It is a difficult task for thehardware designers. When we are dealing with the real-time image data, we don’tknow what we want to do, what we should do is to keep the image information asmuch as its initial data, but we want it a good picture. This operation can improve theimage performance at the source of the system image data and can get more accurateinformation of the scenes.