SAR image tracking based on moment invariants correlation and genetic algorithm

SAR (synthetic aperture radar) image tracking is one of the key technologies of MMW(millimetre wave) imaging missile guider. Due to the restriction of missile-borne imaging conditions such as vibration, rolling and yawing, traditional tracking algorithm with NPROD(normalized production) correlation is difficult to adapt deformations such as rotation and scaling between target image segmented from the previous real-time image and interested area in current realtime image. This paper presents a method to combine moment invariants features with NPROD correlation. Firstly we compute moment invariants features of the previous target image and the current local real-time image respectively. Then we calculate the correlation coefficient of the two groups of feature vectors using NPROD correlation and take it as the function of adaptability degree. Finally we find out the optimal matching position with the GA´s (genetic algorithm) non-ergodic searching mechanism. Experimental result shows that this method not only can solve the problem of deformation image matching but also can greatly improve searching speed and save much matching time with genetic algorithm.