基于强度滤波器的多目标跟踪算法研究

发布时间：2018-02-02 14:57

本文关键词： 多目标跟踪强度滤波器概率假设密度滤波器扩展目标　出处：《西安电子科技大学》2015年硕士论文　论文类型：学位论文

【摘要】：随着多目标跟踪技术的发展,其在国防领域和民用领域的应用也越来越广泛。根据目标产生量测个数的不同,目标可以被建模为点目标和扩展目标。本文结合国家自然科学基金,主要研究强度滤波器及其特例-概率假设密度滤波器在多点目标和多扩展目标跟踪领域中的应用,取得的主要成果如下:1.针对被动多传感器新生未知的情况提出一种基于强度滤波器(Intensity Filter,iFilter)的被动多传感器目标跟踪算法,首先对连续时刻的数据进行关联,得到所有可能的新生目标,并利用量测筛选出真实的目标,完成对目标状态的修正,提高了算法适用范围。2.针对标号多目标强度滤波器(Marked Multitarget Intensity Filter,MMIF)在多点目标跟踪的应用,给出了其序列蒙特卡罗实现方法,验证了该算法在多点目标跟踪中的可行性。3.针对扩展目标新生未知问题,提出一种基于量测数估计的未知新生多扩展目标跟踪算法,计算每个可能的目标对应的量测数,将对应多个量测的目标状态视为真实的目标状态保留下来,扩大了扩展目标的应用范围。4.针对多扩展目标跟踪中量测划分带来的计算量大的问题,提出一种基于迭代映射的多扩展目标跟踪算法,在不需要考虑扩展目标形态的情况下,将扩展目标转化为点目标进行处理,并在此算法的基础上,针对直线运动新生未知的目标提出了一种基于直线拟合的未知新生多扩展目标跟踪算法,提高了多扩展目标的跟踪效率。
[Abstract]:With the development of multi-target tracking technology, it is more and more widely used in the field of national defense and civil affairs. Goals can be modeled as point goals and extended goals. This paper mainly studies the application of intensity filter and its special cases-probability hypothetical density filter in the field of multi-point target and multi-extended target tracking. The main results are as follows: 1. A new intensity filter based on intensity Filter is proposed to deal with the unknown situation of passive multisensor. IFilter-based passive multi-sensor target tracking algorithm firstly correlates the data at continuous time to obtain all possible new targets and screen out the real targets by measuring. Complete the correction of the target state. The application range of the algorithm is improved. 2. For the labeled multi-target strength filter, the marked Multitarget Intensity Filter is improved. MMIF) is applied to multi-point target tracking, and its sequential Monte Carlo implementation method is presented, which verifies the feasibility of the algorithm in multi-point target tracking. 3. In this paper, an unknown new multi-extended target tracking algorithm based on the estimation of measurement numbers is proposed. The corresponding measurements of each possible target are calculated, and the target state corresponding to multiple measurements is preserved as the real target state. The application scope of extended target is expanded. 4. Aiming at the problem of large computation caused by measurement partition in multi-extended target tracking, a multi-extended target tracking algorithm based on iterative mapping is proposed. Without considering the shape of the extended target, the extended target is transformed into a point target for processing, and on the basis of this algorithm. In this paper, a new tracking algorithm based on line fitting for unknown targets with unknown linear motion is proposed, which improves the tracking efficiency of multi-extended targets.
【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】：2015
【分类号】：TN713;TP212.9

【相似文献】