The Track Before Detect Algorithms Based On Hough Transform For Weak Target

With the rapid development of stealth technology in the military field, the radar cross section (RCS) of dim target becomes smaller and smaller, which brings a great challenge to the radar detection system. Track-Before-Detect (TBD) method can effectively detect dim target by multi-frame accumulation with raw data or low-thresholded data. Among all the TBD techniques, the Hough transform based TBD (HT-TBD) method can effectively extract the target trajectory in heavy clutter and hence has important theoretical significance and military value.In this thesis, the HT-TBD technique is investigated and improved, and the main contributions are given as follows:Firstly, in order to reduce the computational and storage burden of the conventional HT-TBD algorithms, a fast ergodic Hough transform (FEHT) algorithm is proposed. It utilizes the motion characteristics of target to establish the validation gate in two adjacent frames and combines all possible measurement pairs to generate possible target parameters. The proposed method can improve the accumulation peak in Hough space, which helps to improve the detection performance. Simulation results show the effectiveness of the proposed method.Secondly, the FEHT method is extended in polar coordinate system. A Fast Polar Hough transform (FPHT) algorithm is proposed with the information of target speed and radar echo frame number. A two-level joint validation gate is established to combine effective data with speed constraint and angle constraint. Simulation results show the effectiveness of the proposed method.Finally, for the problem of detecting dim target in multi-static radar system, a polar HT-TBD based multi-static radar fusion detection algorithm is proposed. The radar echo in single radar is processed separately, and then the estimate result of single radar is sent to the data fusion center. Then, the data fusion center generates the target trajectory by accumulating the same target amplitude. Once the accumulation value exceeds the threshold, a target is announced.Simulation results show that detection performance of the multi-static radar is better than that of single radar.