Research On Digital Image Stabilization In Infrared Homing Seeker Of Single Channel-controlled Air Defense Missile

It is a new application for DIS (Digital Image Stabilization) to the infrared image processing field, in infrared homing seeker of single channel-controlled air defense missile. The disadvantageous effects to target seeking which caused by missile motion, can be eliminated conveniently by the use of computer processor, cooperated with special DIS algorithm. And at the same time, some important motion parameters, such as angle velocity, can be estimated, which help to form accurate control instruction to tracking system. Combined with the motion characteristics of single channel-controlled air defense missile, we study the imaging theoretical model and several key technologies in MDIS (Missile Digital Image Stabilization) in this thesis.Firstly, we discuss the traits of missile and targets, and discuss other related problems, which are concretely as follows, the characteristics of targets and environment in battlefield, homing detect system, homing control system, and some important issues in control system. Based on the discussion above, the holistic searching frame of MDIS is established. Through which, the whole MDIS problem is divided into six independent modules, and the relations between each modules are indicated subsequently.Secondly, three motion models are established according to the imaging laws in seeker and motion characteristics of target and missile, which orderly are GMSM (General Motion Space Model that target related to rotated missile body), PMSM (Motion Space Model that contain Precession in missile body), and RMIM (target-missile Relative Motion Model in Imaging surface).Thirdly, based upon RMIM, two sets of motion parameters estimation algorithm are raised, which respectively are EBPA (estimation algorithm based on polar coordinates analyzing) and EBGA (estimation algorithm based on geometry characteristic analyzing). Under the condition of point target background, the rotation parameter and precession parameter can be estimated effectively. The theoretical analysis shows that there exists no algorithm error in EBGA. Subsequently, an integrated MDIS workflow under point target background is set up, which is based on EBGA.And fourthly, for the image sequences in which contain simple background scenery, another MDIS workflow is set up after the study of DIS algorithm which based on RMIM and features matching. According to RMIM, a new feature points matching approach is presented, which is based on crossed image blocks. This approach can extract global feature points which are separated homogeneously. When we estimate motion parameters,optimization is realized by simulated annealing. And two kinds of filter algorithms are designed respectively to smooth the motion parameters, which are Kalman-based filter algorithm and forecast filter algorithm based on unitary liner regression.Finally, three simulation systems are developed to validate the efficiency of theories raised above. They are relative motion orbit simulation system for point target, MDIS simulation system for point target, and MDIS simulation system for general image sequences which include background scene. The simulation results show the studies in this thesis are feasible.