The Research And Implementation Of A Morris Water Maze Video Analysis System Based On Object Detection And Tracking

Moving object detection and tracking based on video sequences is one of the key techniques in the field of computer video analysis, which combines advanced technologies in digital image processing, computer graphics, pattern recognition, artificial intelligence and other relative fields. It has very important practical value and wide developmental prosperity in military surveillance, intelligent transportation, medical diagnosis, industrial product detection and so on. This paper mainly focuses on analyzing and processing videos with moving objects in Morris water maze medical experiments by using techniques in moving object detection and tracking fields, which has great theoretical meaning and practical value for propelling the research of learning and memory in medical studies.In-depth researches on moving object detection and traking based on video sequences have been made in this paper, and based on these researches a Morris water maze video analysis system has been developed. The main research can be summarized as follows:An avi video frame grabbing algorithm is designed to convert avi videos to continuous bitmap image sequences. Filtering algorithms base on space, frequency, wavelet and time have been introduced to reduce noises caused by environmental and lighting conditions in the video image sequences. These methods not only reduce noises, but also enhance the contrast of the image sequences, which solve the identifiability of the underlit images with good performance and therefore provide a sound guarantee for subsequent object detection and tracking procedures.In terms of moving object detection, some classical detecting algorithms in the field of object detecting are introduced, including frame subtraction, background subtraction and kim after a systematical study of the theory of object detection. All those algorithms are analyzed and their advantages and disadvantages are included. This paper adopts the background subtraction algorithm to detect the swimming rats in the Morris water maze videos according to the research requirement of the study. In order to segment the rat from the maze effectively, an improved Otsu algorithm based on gray stretch is designed. Experiments show that the improved algorithm can achieve an ideal result both in efficiency and accuracy.In terms of moving object tracking, this paper attaches great importance to template matching based tracking algorithm, Kalman filter based prediction tracking algorithm and CamShift tracking algorithm after an thorough understanding of the common algorithms in the field of moving object tracking. Finally, a minimum area searching algorithm based on template matching is designed according to the own features of the system. This algorithm firstly calculates a minimum searching area, then predicts the possible position the rat may be in the next frame based on the coordinates obtained in the previous frame. Experiments show that the algorithm can precisely track the rat in the Morris water maze.Lastly, a Morris water maze video analysis system based on moving object detection and tracking is designed and realized. Repeated experiments show that the system can detect and track the swimming rat in Morris water maze effectively with a satisfactory performance which brings practical value to the research on the effectiveness of the new drugs for the treatment of dementia caused by diabetes mellitus.