Based On The Image Of The Target Tracking Feature Extraction Method

With further requirement of the advanced technology in military tasks, the technology of imaging-based target tracking ,which is an important development direction of target tracking, gets quick of development, and is gradually adopted in the system of precision guided weapons. Feature point extraction and tracking method is an important contents to studying imaging-based target tracking, and avails to the research on target identification,maneuver detection and so on.In this paper, feature point extraction and tracking problem are put together in the research. Based on study of Gabor wavelet transformation based feature point extraction algorithm, a new feature point tracking scheme is designed, which has intelligent effects when tracking a set of feature points. The main achievements are summarized as follows:1.Based on Deeply studying the feature of the Gabor transformation, a characteristic vector for describing points' feature is defined with the gabor transformation to realize feature point extraction.2.The merits of existing techniques for tracking feature point are considered in designing a new algorithm of tracking feature points. PDAF, image matching, EKF and e.t.c. is used in this scheme . The simple affine transformation is used to describe the motion of feature points, and an PDAF is employed to obtain the estimate of the state vector; A matching method based on Gabor wavelet transformation is employed to find the corresponding point; An EKF is employed to process the temporal information; A scheme is designed to include new feature points. The simulation results show this scheme's improvement.3.A new matching based on Gabor wavelet transformation is designed on the analysis existing matching method. The simulation results have verified the method's validity.Therefore, the research results in this paper would be useful for the technologyof imaging-based target tracking.