Research On Restoration Of Images Collected By Reconnaissance System In Near-Space

During the past years, near-space platforms become the researching hotspot of major military powers. Because of its the special place which between air and space, the establishment of the military reconnaissance systems in this platform not only has its unique advantages, but also it could work together with the aerospace platform, which opened up a wider space for the military reconnaissance operation. And it also provides more reliable information for many post-military tasks.Reconnaissance in near-space is a systematic work. The vehicles designed according to features of near-space platforms carry the the image sensor and detect the target region of interest image,and the images return the ground control center collected through wireless channel, and do image-analysis according to different application requirements, find the reasons of image-degradation, and then we can restore them.In this paper,we research on compression and restoration of interested target area images collected by reconnaissance system of near-space, many problems of this system are to be resolved. We have done following work:The first is near space platforms of analytical work. Near-space platform as a new platform, its features are what people most want to know, and is also the primary task. Only by understanding the space, could we better design of aircraft and aircraft carrying various types of sensors. In this paper, we do a comprehensive analysis of the approach and differences between near-space and satellite platform.Subsequently, do the two parts of image restoration work for different causes of degradation.Firstly, it's image restoration of compression distortion images.The volume of images data collecting in near-space is huge, in order to both save storage space and reducing the transmission channel capacity, image compression is necessary.The current compression standard is variety, we choose the standard JPEG. It's simple, mature, and easy to implement.Not only it can still image compression, but also to frame as a unit for image compression, so it is very suitable for the new space exploration and research. In this paper, we design a near-space simulation image compression system.Image compression is imperative, and the corresponding image degradation is unavoidable.As compression system used in the JPEG compression standard, when we use a greater compression ratio, the compression distortion mainly react to blocking artifacts, to enhance the image quality, the study corresponding image restoration technology is very necessary. The main reason of blocking-artifacts is after image-quantify the uneven distribution of high-frequency information between blocks. Wavelet transform method has the good nature of the image frequency domain analysis to the wavelet transform domain process. And in this paper, we use wavelet transform method to restore the images and achieve better results.Second, it's the stereo image restoration technology research. In some deep-level targets such as identification, the battlefield application, stereo images can provide significant help to facilitate the mission. So restoring the goal of the third-dimensional space of information and researching resumption of three-dimensional technology is very necessary.In this paper, we based on the stereo visual principle, with two images collected from different angles to achieve recovery tasks. The major study is on the imagepair matching which is the key technology of stereo image restoration. With the result of matches we resume the stereo information of the images.