基于空间栅格划分的协同探测系统检测算法研究

发布时间：2019-05-22 18:28

【摘要】：现代科技战争中,采用协同探测系统是应对低截获目标的有效手段。通过在不同维度上采用空间分集、频率分集等技术可以有效的克服基于后向散射特性的隐身目标探测难的问题。现阶段采用协同探测系统检测与跟踪目标仍停留在理论探索阶段,如何在实际应用中实现该技术仍是一个难题。在理论研究阶段,很多学者对协同探测系统信号级联合检测技术做了大量研究,然而这些研究大多建立多站点雷达的各通道检测单元已经理想配准的假设之上。这个假设在实际应用中往往是不成立的。另外,协同探测系统的检测算法初步检测产生的结果往往由于多通道波束交叠产生很多虚假目标。在实际的工程应用中,需要对大型区域进行监视,考虑计算量因素,只能对空间进行粗略的栅格划分。出现目标后,则希望根据目标信息通过细粒度的栅格划分提高检测和跟踪精度。这种情况下如何实现目标跟踪、划定监视区域、确定栅格宽度都是需要解决的实际问题。本文针对以上问题,提出了一套包含信号预处理、目标检测、位置估计、目标跟踪的基于空间栅格划分的协同探测系统检测算法,具体如下:(1)针对协同探测系统多通道联合检测空时失配的问题,提出了基于数据对齐的检测单元配准准则。该准则在空间栅格宽度较小的场景中检测与位置估计精度均较高,计算量较小。(2)针对基于数据对齐的检测单元配准准则在空间栅格宽度较大的场景检测与位置估计性能下降的问题,提出了基于极值搜索的检测单元配准准则。该准则检测性能随栅格宽度增加导致的衰减极小,在栅格宽度较大场景中位置估计精度优于数据对齐准则。(3)针对多个通道检测单元重叠产生虚假目标的问题,研究了基于逻辑判决的虚假目标消除算法和基于Clean的虚假目标凝聚技术。在栅格宽度不远大于距离单元的情况下,两个方法的联合使用能够有效消除虚假目标。(4)针对大型监视区域难以通过离线空间栅格进行检测跟踪的问题,提出了基于动态空间栅格划分的目标检测跟踪算法。该方法在保持跟踪精度不变的条件下,降低了计算量。文中提出的计算方法都通过仿真进行了验证,证明了基于空间栅格划分的协同探测系统检测算法的有效性。
[Abstract]:In modern science and technology war, cooperative detection system is an effective means to deal with low intercepted targets. By using spatial diversity in different dimensions, frequency diversity and other techniques can effectively overcome the problem of stealth target detection based on backward scattering characteristics. At present, the detection and tracking of targets by collaborative detection system is still in the stage of theoretical exploration, and how to realize this technology in practical application is still a difficult problem. In the theoretical research stage, many scholars have done a lot of research on the signal level joint detection technology of collaborative detection system. However, most of these studies have established the assumption that each channel detection unit of multi-station radar has been ideal registration. This assumption is often untenable in practical application. In addition, the preliminary detection results of cooperative detection system often produce a lot of false targets due to the overlap of multi-channel beams. In practical engineering applications, it is necessary to monitor large areas, and only rough grid division of space can be carried out considering the calculation factor. After the target appears, it is hoped that the detection and tracking accuracy can be improved by fine-granularity raster partition according to the target information. In this case, how to achieve target tracking, delineate monitoring area and determine grid width are practical problems to be solved. In order to solve the above problems, a set of cooperative detection system detection algorithm based on spatial grid partition is proposed, which includes signal preprocessing, target detection, position estimation and target tracking. The details are as follows: (1) aiming at the problem of space-time mismatch in multi-channel joint detection of collaborative detection system, a registration criterion of detection unit based on data alignment is proposed. The detection and position estimation accuracy of this criterion is higher in the scene with small spatial grid width. The computation is small. (2) aiming at the problem that the detection unit registration criterion based on data alignment degrades the performance of detection and position estimation in the scene with large spatial grid width, a detection unit registration criterion based on extreme value search is proposed. The attenuation caused by the increase of grid width is very small, and the accuracy of position estimation is better than that of data alignment criterion in the scene with large grid width. (3) aiming at the problem of false targets caused by the overlap of multiple channel detection units, The false target elimination algorithm based on logical decision and the false target condensation technology based on Clean are studied. When the width of the grid is not much larger than the distance unit, the combined use of the two methods can effectively eliminate the false target. (4) aiming at the problem that it is difficult to detect and track the large monitoring area through the offline space grid, A target detection and tracking algorithm based on dynamic spatial grid partition is proposed. This method reduces the amount of computation while keeping the tracking accuracy unchanged. The calculation methods proposed in this paper are verified by simulation, and the effectiveness of the detection algorithm based on spatial grid partition is proved.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TN957.51

【参考文献】