光学学报, 2017, 37 (7): 0728003, 网络出版: 2017-07-10
基于改进豪斯多夫距离的扩展目标形态估计评估
Shape Estimation Evaluation of Extended Objects Based on Modified Hausdorff Distance
遥感 性能评估 扩展目标 形态估计 改进的豪斯多夫距离 remote sensing performance evaluation extended object shape estimation modified Hausdorff distance
摘要
为了验证高精度传感器量测下某种扩展目标跟踪算法的有效性,往往需要与其它算法进行对比,并评估其估计性能。与传统的点目标不同,扩展目标跟踪的主要任务不仅仅是要估计出目标的运动状态,更重要的是要对其扩展形态进行精确估计。因此,对扩展目标的形态估计性能评估有着迫切的需求。针对基于星凸形和支撑函数这两种具有代表性的扩展目标模型,考虑到其所具有的不同的形态参数描述方式,提出了一种具有不同数学形式的改进豪斯多夫距离来解决此问题。仿真实验表明,提出的改进豪斯多夫距离能够作为一种有效的度量准则来对扩展目标的形态估计性能进行有效评估。
Abstract
In order to verify the validity of extended objects tracking algorithm under high precision sensors measurement, it is often necessary to compare with other algorithms and evaluate its estimated performance. Unlike classical point objects, the main task of extended objects tracking is not only to estimate the motion state of the objects, but also more importantly to accurately estimate their shape. As a result, there is an urgent need to the shape estimation performance evaluation of the extended objects. Aiming at the two kinds of representative extended target models based on star-convex and support functions, considering their different morphological parameters, an improved Hausdorff distance with different mathematical forms is proposed. Simulation results demonstrate that the modified Hausdorff distance can be used as an effective metrics to evaluate the shape estimation performance of the extended targets effectively.
孙力帆, 张森, 冀保峰, 普杰信. 基于改进豪斯多夫距离的扩展目标形态估计评估[J]. 光学学报, 2017, 37(7): 0728003. Sun Lifan, Zhang Sen, Ji Baofeng, Pu Jiexin. Shape Estimation Evaluation of Extended Objects Based on Modified Hausdorff Distance[J]. Acta Optica Sinica, 2017, 37(7): 0728003.