Research On Moving Target Tracking Method In Image Sequences Under Complex Background

With the development of infrared imaging technique, infrared imaging is widely applied in object detection and object recognition and so on. Detection of moving object in image sequences is the process that separates moving object from static background. It has been an important role in the field of moving object detection.An algorithm based on template matching is designed for moving object detection in this thesis. The detection algorithm is divided into two steps: pre-difference detection and domain blocks detection. Pre-difference detection is used for detecting moving objects in the current image. Once there exist moving objects they will be segmented. Then the template matching is used in detecting the domain blocks in domain blocks detection. Finally, moving objects tracks are obtained. An inter-frame difference method is improved in the thesis and applied in the pre-difference detection. Then a segmentation result is obtained by means of basic methods of morphology based on the inter-frame difference method. An template of domain blocks detection is advanced and proposed in the thesis, then the matching is performed based on the result of pre-difference detection. The results show that the pre-difference detection and domain blocks detection have been combined effectively.