Aiming at the large cost of calculating variable bandwidth kernel particle filter and the high complexity of its algorithm,a self-adjusting kernel function particle filter is presented. Kernel density estimation is fa...Aiming at the large cost of calculating variable bandwidth kernel particle filter and the high complexity of its algorithm,a self-adjusting kernel function particle filter is presented. Kernel density estimation is facilitated to iterate and obtain new particle set. And the standard deviation of particle is introduced in the kernel bandwidth. According to the characteristics of particle distribution,the bandwidth is dynamically adjusted,and the particle distribution can thus be more close to the posterior probability density model of the system. Meanwhile,the kernel density is used to estimate the weight of updating particle and the system state. The simulation results show the feasibility and effectiveness of the proposed algorithm.展开更多
Mean Shift是一种基于特征的对目标实现快速跟踪的算法,传统的Mean Shift算法由于跟踪中物体的尺度变化会使跟踪偏离目标乃至跟踪失败,并且原有的自适应地对跟踪窗宽的调整,是基于对核窗宽的改变来得到的。在"固定跟踪窗宽—改变...Mean Shift是一种基于特征的对目标实现快速跟踪的算法,传统的Mean Shift算法由于跟踪中物体的尺度变化会使跟踪偏离目标乃至跟踪失败,并且原有的自适应地对跟踪窗宽的调整,是基于对核窗宽的改变来得到的。在"固定跟踪窗宽—改变核窗宽"的基础上对目标进行跟踪,对目标空间定位精度进行了评估与分析,通过实验结果表明改变核函数参数能改善目标跟踪的精度。展开更多
基金Supported by the National Natural Science Foundation of China(60972059)the General Project of Science and Technology of Xuzhou City(XM12B002)
文摘Aiming at the large cost of calculating variable bandwidth kernel particle filter and the high complexity of its algorithm,a self-adjusting kernel function particle filter is presented. Kernel density estimation is facilitated to iterate and obtain new particle set. And the standard deviation of particle is introduced in the kernel bandwidth. According to the characteristics of particle distribution,the bandwidth is dynamically adjusted,and the particle distribution can thus be more close to the posterior probability density model of the system. Meanwhile,the kernel density is used to estimate the weight of updating particle and the system state. The simulation results show the feasibility and effectiveness of the proposed algorithm.
