基于动态规划的微弱目标检测前跟踪算法研究

发布时间：2019-06-19 21:07

【摘要】：随着高速目标和隐身目标的出现,微弱目标的检测与跟踪受到了广泛的关注。检测前跟踪算法是一种高效的实现微弱目标检测与跟踪的手段。当跟踪低信噪比目标时,它通过连续处理多帧数据沿可能的目标航迹积累能量来同时实现目标的检测与跟踪。动态规划算法是一类多阶段决策优化问题,它是检测前跟踪方法中一种高效的检测与跟踪微弱目标的方法,因此它成了国内外研究的热点。本文对基于动态规划的检测前跟踪算法进行了系统的研究,所取得的主要研究成果如下：1.对传统的基于动态规划的检测前跟踪算法进行了研究。首先介绍了动态规划的基本原理,然后以高斯噪声下匀速运动的点目标为例研究了基于动态规划的检测前跟踪算法的实现流程,并提出了一种重叠航迹法与统计方向直方图相结合的去除伪航迹的新算法,最后详细分析了动态规划检测前跟踪算法的性能。2.针对传统动态规划检测前跟踪算法的缺点,提出了三种改进的动态规划检测前跟踪算法。针对计算量大的缺点,提出了两级检测门限算法,并且为了保证算法的性能,在动态规划算法递归积累过程中加入了值函数转移步骤,结果表明,在很小的性能损失下可以大幅的减小计算量；对于复杂环境下的微弱目标,提出了一种利用复数据的相位信息和杂波图信息的复似然比检测前跟踪算法。由于相位信息和杂波图信息的使用,该算法可以有效的从非均匀强杂波环境中检测出目标,与此同时,与幅度似然比算法相比,也大大减小了计算量；由于雷达信号处理本身的特点,提出了一种利用多普勒信息的检测前跟踪算法。该算法充分利用多普勒信息减小搜索范围,不仅减小了计算量,而且减小了能量扩散及强杂波或噪声点的影响。与光学图像中微弱目标的检测跟踪相比,特别是在复杂的环境下该算法可以实现更优的检测与跟踪性能。3.研究了基于动态规划的多目标检测前跟踪算法。首先介绍了多目标检测前跟踪算法的理论模型并分析了多目标检测前跟踪实现的难点；然后通过引入广义检测程序,在前两种算法的基础上,提出了一种新的利用多目标结构的多目标检测前跟踪算法。这三种算法其实质都是如何将多目标的检测跟踪转化成多个单目标的检测与跟踪；最后通过仿真对三种算法的性能进行了对比分析,结果表明本文新提出的算法性能最佳,并用实测数据验证了该算法的有效性。
[Abstract]:With the emergence of high-speed targets and stealth targets, the detection and tracking of weak targets has received extensive attention. Pre-detection tracking algorithm is an efficient means to achieve weak target detection and tracking. When tracking low signal-to-noise ratio (SNR) targets, it accumulates energy along possible target tracks by continuously processing multiple frames of data to achieve target detection and tracking at the same time. Dynamic programming algorithm is a kind of multi-stage decision optimization problem. It is an efficient method to detect and track weak targets in the pre-detection tracking method, so it has become a hot research topic at home and abroad. In this paper, the pre-detection tracking algorithm based on dynamic programming is studied systematically, and the main research results are as follows: 1. The traditional pre-detection tracking algorithm based on dynamic programming is studied. This paper first introduces the basic principle of dynamic programming, then takes the point target moving at uniform speed under Gaussian noise as an example to study the realization flow of the pre-detection tracking algorithm based on dynamic programming, and proposes a new algorithm for removing false tracks based on overlapping track method and statistical direction histogram. Finally, the performance of the tracking algorithm before dynamic programming detection is analyzed in detail. 2. In view of the shortcomings of the traditional dynamic programming pre-detection tracking algorithm, three improved dynamic programming pre-detection tracking algorithms are proposed. Aiming at the disadvantage of large amount of computation, a two-level detection threshold algorithm is proposed, and in order to ensure the performance of the algorithm, the value function transfer step is added to the process of recurrent accumulation of dynamic programming algorithm. The results show that the computation can be greatly reduced under very small performance loss. For weak targets in complex environment, a complex likelihood ratio tracking algorithm using complex data phase information and hash graph information is proposed. Due to the use of phase information and hash map information, this algorithm can effectively detect targets from non-uniform strong hash environment, at the same time, compared with amplitude likelihood ratio algorithm, it also greatly reduces the amount of computation. Because of the characteristics of radar signal processing itself, a pre-detection tracking algorithm using Doppler information is proposed. The algorithm makes full use of Doppler information to reduce the search range, which not only reduces the amount of computation, but also reduces the influence of energy diffusion and strong clutters or noise points. Compared with the detection and tracking of weak targets in optical images, especially in complex environments, the algorithm can achieve better detection and tracking performance. The pre-detection tracking algorithm of multi-target detection based on dynamic programming is studied. Firstly, the theoretical model of multi-target detection pre-tracking algorithm is introduced and the difficulties of multi-target detection pre-tracking are analyzed. Then, by introducing the generalized detection program, a new multi-target detection pre-tracking algorithm using multi-target structure is proposed on the basis of the first two algorithms. The essence of these three algorithms is how to transform the detection and tracking of multiple targets into the detection and tracking of multiple single targets. Finally, the performance of the three algorithms is compared and analyzed by simulation, the results show that the new algorithm has the best performance, and the effectiveness of the algorithm is verified by the measured data.
【学位授予单位】：西安电子科技大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN911.23

【相似文献】