Research On Novel Detection Algorithms For Moving Infrared Dim Small Targets

The detection and tracking of moving dim target is a focus in computer vision and image processing, and is also the core of infrared detection and surveillance system, especially in the condition of complicated background and far distance. The problem of detection and tracking of dim point target comes from application of long-range surveillance system, such as wide FOV (field of view) telescope and SBIRS (Space-based infrared system) and other surveillance system. The dim target usually hides in the constructed background clutter and heavy noise environment. In recent years, researchers pay more and more attention on the detection and tracking of dim target in IR image sequence. The performance of the algorithm is very important to the effective distance and intelligent extent of the IR surveillance system.The purpose of the paper is to propose novel dim target detection and tracking algorithms in low signal to noise(SNR) condition with little computing cost and Robust characteristic. The work of the paper provides theory foundation and experience support for the real-time implementation of the National Natural Science Foundation of PR China and the National emphatic pre-research project of PR China. The work of the paper focuses on two aspects. First of them is the research work of dim moving target detection technology in IR image sequence; and the second is the research work of dim moving target tracking algorithm and target trajectory confirmation and detection algorithms'performance analysis in IR image sequence. The work of the paper provides technology support of developing the infrared detection and surveillance application system and other corresponding research work.Analysis the background characteristic and the method of frames difference measuring of IR image sequence. Describe the common model of the IR image background prediction. Discuss the theory of IR image background suppression based on the background suppression and dissertate six kinds of background suppression techniques in detail. Analysis the applicability of those background suppression techniques depend on the experimental results.Describe the common theory of one dimension and two dimension Wiener filters. Develop a novel dim target detection algorithm of IR image by combining the two dimension Wiener filter with the temporal prediction filter named three-dimensional Spatial-Temporal Adaptive Wiener Prediction Filter (3DSTAWPF) method. And apply it in the single spectrum dim moving target detection and tracking system which combined high speed scanning detector with staring detector. These testify its practicability and validity.Analyze the filtering characteristic of the anisotropic diffusion method. Develop a novel Double Threshold Partial Differential Equation method based on the anisotropic diffusion, and use it to target detection in IR images creatively. By comparison of other algorithm, the experimental results demonstrate that the new method can suppress the background effectively and is better than other algorithm, especially in constructive texture suppression of background and keep the target signal steadily. The experimental results testify its practicability and validity in target detection.Three-dimensional directional filter is a classical method of target detection. It reinforces the detectability greatly by accumulating the targets energy. In the classic model of three-dimensional directional filter, the target energy accumulating ability depends on the frames of the processed frames monotonously. The paper develop a novel three dimensional Wide-to-Exact Search Double Directional Filter. It uses two different search precision to search the whole state space and corresponding state space, and gets a good compromise of compute cost and detection performance. The target energy accumulating ability of the new algorithm is only depend on the length of integrating time, and independent of the integrating moment.Develop the target tracking method in the single spectrum dim moving target detection and tracking system which combined high speed scanning detector with staring detector by using common threshold technique, depending on the continuity and orientation of the target motivation, analyzing imaging characteristic of scanning detector and staring detector, using least mean square prediction technique.Analyze the detection probability and false alarm probability theoretically and buildup models and make experiments by using target detection algorithm performance evaluation method and corresponding precondition and hypothesis, considering target/background statistic characteristic.