Research On Track-before-detect Algorithm Based On Dynamic Neighborhood

In modern battlefields,especially in urban environments,Track Before Detect(TBD)technology for weak targets(especially stealth targets or covert targets)has drawn increasing attention.The basic idea of TBD is to process the continuous frame data together,track the target before detecting the target,so as to realize the non-coherent accumulation of multi-frame data and improve the detection performance.In this paper,a TBD algorithm based on Dynamic Programming(DP)without any a priori limitation on the target motion is selected to study the detection and tracking of indoor personnel through the wall.Obtain the following main research results:1.The key parameters affecting the DP-TBD algorithm are analyzed and the applicable conditions of the traditional DP-TBD algorithm are given.The ideal characteristics of dynamic programming applied to the optimization problem are analyzed.The basic recurrence equation of DP-TBD algorithm is given.The DP-TBD algorithm for detecting and tracking indoor targets is studied.The algorithm uses the image amplitude as an objective function to effectively track and monitor the indoor targets whose signal-to-noise ratio is greater than-10 dB.2.Aiming at the disadvantage of large computational load of DP-TBD algorithm,NS-TPDP-TBD algorithm is proposed to detect and track regular particle targets.The criteria for selecting the first-level threshold in TPDP-TBD algorithm are given.The NS-TPDP-TBD algorithm,which is suitable for detecting and tracking indoor rule particle points,is proposed.The algorithm introduces target prior information and improves the DP-TBD algorithm so that the search times are no longer related to the image pixel size,Threshold detection may produce a number of suspected target trajectory,thereby reducing the amount of search algorithm to improve the timeliness of the algorithm.The simulation results show that the NS-TPDP-TBD algorithm is more efficient and better performance than the DP-TBD algorithm and insensitive to the indoor personnel's flicker problem.3.Aiming at the practical problems of NS-TPDP-TBD algorithm,an ADNS-TPDP-TBD algorithm is proposed for irregular extended target detection and tracking.The practical problems faced by the NS-TPDP-TBD algorithm are analyzed.An improved algorithm,called the ADNS-TPDP-TBD algorithm,is proposed to detect the indoor irregular expansion target.The algorithm introduces equivalent neighborhoods and analyzes the dynamic expansion of neighborhoods to avoid the strong clutter interference caused by multipath reflections and to solve the problem of irregular expansion of personnel targets and changes of moving speed.Simulation and measured data show that the ADNS-TPDP-TBD algorithm can achieve multi-person tracking,can adapt to multi-path clutter environment,is not sensitive to the irregular expansion of personnel target and speed variation,and is more suitable for practical application than NS-TPDP-TBD algorithm,With good practical value.