Dynamic programming network for point target detection

Abstract To improve the efficiency of dim point target detection based on dynamic programming (DP), this paper proposes a multi-frame method DP ring network (DPRN). In proposed method, first, trajectory is approximated using piecewise linear function. The velocity space partition (VSP-DP) used to accumulate merit functions each segment avoid function diffusion in different spaces. addition, matching (VSM-DP) employed realize state transition between adjacent segments. Then, VSP-DP and VSM-DP are construct (DPN). Second, suppress further, sequential reverse DPNs connected head-to-tail manner form DPRN, DPRN obtained by averaging DPNs. Finally, tracking extreme points DPRN. simulation analysis results show that combines advantages high probability high-order execution first-order DP. suitable for radars infrared searching systems.