Research On Information Processing And Experiment Of Strap-Down Imaging Guidance

The whole strap-down imaging guidance system contains the acquisition of image, the detection and tracking of moving target, the guidance law, and the attitude control. For the practical strap-down imaging guidance simulation system, some problems concerned are studied, including image acquisition, image preprocessing, moving target detection. And the algorithm implementation is given.First, the hardware architecture of the image acquisition system is introduced, and the program is given to realize adjustable frame rate and exposal time, and image saving.Then, the thesis introduces the principle, characteristic of image preprocessing algorithms. Particularly, noise suppression and gray transform algorithms are introduced in detail.It is often needed to design 3D spatio-temporal filter bank in frequency domain for the implementation of the conventional moving target detection algorithms, where there exist such problems as many parameters to choose, difficulty in filter design. While in the thesis, a novel moving target detection algorithm in 2D frequency domain is proposed. By decomposing the dynamic images into row and column sequences, the detection problem in 3D frequency domain is converted into detection problems in 2D frequency domain, which leads to easy design of filters. Then an adaptive filter is designed to extract the motion energy and to eliminate the frequency components of static background and noises, and the moving targets can be detected effectively.At last, moving target detection algorithm are implemented to test moving target detection effects under the conditions of noise disturbance, sensor dithering and target sheltered. The results justify the robustness of the proposed methods against image noise and registration errors.