A Maximum Fuzzy Entropy Gaussian Clustering Based Tracking Technology For Multiple Moving Point Targets In Image Sequences

With the development of missile technology and infrared technology advances, more and more attention is paid to research on the infrared imaging guidance technology. The technology of dim infrared target tracking is one of the key technologies in the system of infrared image.Around the problem of the tracking for dim targets in low SNR images, this paper introduces related theory,derives the maximum fuzzy entropy Gaussian clustering whose clustering effect better than the traditional FCM, mainly researches on the technology based on the maximum fuzzy entropy Gaussian clustering for tracking Multiple Cross-moving Point Targets . At first, we get all targets'reliable initial information based on TBD detection method in image sequences. And in order to improve the real time quality of this algorithm, we adopt the single-frame detection technology in the tracking stage, then in the tracking window, we use the degrees of membership between the measurements and the predicted locations of the targets which are obtained through maximum fuzzy entropy Gaussian clustering as weight to fuse information to realize the update of targets'states, then we use Kalman filter to predict the next state.Experimental results demonstrate that this technology can accurately track dim point Multi-target in strong clutter environment, and ensures good real-time, has a certain applied value.