A novel HT-TBD detection approach for near-space target

Near-space target detection is usually difficult for radar due to this kind of target's characteristics of hypersonic velocity and weak echoes. A novel detection approach is proposed by employing Hough transform track-before-detect (HT-TBD) technique based on grading and dimensionality reduction mechanism. Firstly, in order to solve huge computational load problem of traditional three-dimensional Hough transform, three-dimensional points are mapped into range-time plane, azimuth-time plane and elevation-time plane successively for three corresponding two-dimensional Hough transforms. Then, a new double integration means of noncoherent integration and binary integration is adopted in parameter space to select points in every stage orderly, which could make full use of energy information and decrease impact from strong interference at the same time. Finally, motion constraints and trajectory merging are introduced to further reduce detected false trajectories after sequence association for detected points. Experimental results demonstrate the effectiveness of proposed approach.