Study On Multispectral Detection System Of Near Space Target

Near space is the space area scope from altitude 20km to 100km. The near space belongs to neither the aviation region nor the space category, it has no the national boundaries. Therefore, many world powers are stepping up the development and utilization of the near space, make great efforts to compete for dominance in the near space. Now many detection equipments, such as stratospheric airships, floating balloon, and unmanned aerial vehicles etc have been developed by many powers. It results in many countries in the world including China are monitored long time.There have no ground-based detection equipment for detecting near space target in the world now. A passive ground-based optical detection system of the near space target is proposed, the system solution includes many spectral band TDI sensors, for example, the visible sensor and the infrared sensor. It mainly used to detect low-dynamic aircraft target in near space.The system mainly includes mechanical scan module, optic imaging module, image transfers module and image processing module. When the system works, servo system controls torque motor work and turntable rotation. At the same time, the encoder provides the pulse signal to the visible camera and the infrared camera, the pulse triggers cameras work. The image data is transferred to image acquisition module through the image transfers system, and then the image data is transferred to the main computer. It is used to process and display in the computer.In the image processing, the keystone is the feature point registration algorithm, Harris extraction algorithm and SUSAN extraction algorithm are described, the coarse matching and the fine matching methods including the partial correlation coefficient and the RANSAC algorithm etc are given in the paper. Finally, experimental results are given. Image fusion contains a lot of efficient and feasible methods. The simple image fusion algorithm and the transform domain-based image fusion algorithm are described in the paper. The Laplacian Pyramid decomposition algorithm is analyzed carefully. The corresponding experimental results are given in the experiment part. Wavelet transform image fusion method is one of the most common methods. Base on the conventional fusion method, an optimized algorithm of wavelet transform fusion rule is given, and the corresponding experimental results are given in the last part.Matlab and VC++ are used to program experiment code, in the course of the test, a lot of material parameters are given, and the feasibility of the system is preliminary verified.