Study On Detection And Tracking Dim Point Moving Target From Image Sequences

Since the target detection and tracking system from optical signal deals with 3D space-time signal and is a kind of passive detection systems, comparing to other target detection systems, it has a lot of features, such as concealment, anti-jamming and high precision tracking performance. At present, the technology for target detection, acquisition, tracking and pointing from optical signal has been one of the key technologies in modern combat system, and has been used in many advanced weapons.Recent years, with the improving requirements from the military and others, in which the targets are required to be detected as soon as possible and as far as possible, there are a great of research interests in studying of the detection and tracking technology of dim point moving target from image sequence. Two important problems have to be resolved in this area, one is how to detect and track dim point moving target in highly cluttered image sequences; the other is how to realize it in real time.The main contribution of the dissertation is the whole studying of the detection and tracking technologies of dim point moving target from image sequence. In this dissertation, the research trends for the problem have been introduced; the theoretical model for the problem has been built. Also in this dissertation, the problem to detect and track dim point moving target from image sequence is divided into three parts, the image pretreatment, the target detection and the target tracing. Every part has been described in detail and a lot of theoretical and experimental results have been analyzed. Several valuable and important results which bring forth new ideas are achieved and listed as the following:1. A new method to compensate moving background based on robust iterative reweighed least squares (RIRLS) was suggested. The camera moving parameters were firstly estimated by using optical flow model and RIRLS method. And then, moving background image was compensated under these parameters. The robust ability to anti-noise and the real-time calculation ability are the main advantages of the method.A method that projected 3D space-time signal to 2D space signal to integrate images for detecting dim point moving target from image sequences was suggested. By the projection transform, the search paths were drastically reduced while the highly detection2. performance was hold.3. A few of dim point moving target detection algorithms from image sequence were proposed. That was the algorithm based on hard-decision integration, the algorithm based on soft-decision integration, the algorithm based on time difference integration and the algorithm based on probability integration. The theoretical and Monte Carlo experimental performances were also showed in this dissertation.4. Based on the basic idea of probability data associating filter (PDAF), a new PDAF algorithm to tracking dim point moving target from images sequence was derived. Theoretical and experimental results showed that the method could keep the highly tracking precision as well as the real-time tracking ability.5. Based on the idea of multi-resolution method, a scheme to detect and track dim point moving target from image sequence was proposed. The scheme can detect and track target, which size wasn't one pixel and which velocity wasn't one pixel per frame...