论文部分内容阅读
多目标跟踪的主要问题是有效区域,即窗口的定义,数据联想和状态预测.窗口定义了一个测量向量的有效区域,它是用高斯、矩形和椭圆函数来定义的.数据联想是一个复杂的任务.许多有关快速和复杂的技术已经报道过.Barshalon等人提出了一个概念,对单目标用概率数据联想,而对多目标,用联合概率数据联想.这些算法在计算上是复杂的,并且在仿真几个目标的时候,实时完成要求庞大的数据吞吐量.
The main problems of multi-target tracking are the valid regions, namely the definition of window, the data association and the state prediction. The window defines the effective area of a measurement vector, which is defined by Gaussian, rectangular and elliptic functions. Data Association is a complex Many fast and complex techniques have been reported.Barshalon et al. Proposed a concept associative with probabilistic data for a single target and associative data with a joint probability for multiple targets.These algorithms are computationally complex and While simulating several goals, it requires massive data throughput in real time.