基于GPU加速的粒子滤波多说话人跟踪算法及其应用 |
| |
引用本文: | 曹洁,黄开杰,王进花. 基于GPU加速的粒子滤波多说话人跟踪算法及其应用[J]. 计算机应用研究, 2019, 36(3) |
| |
作者姓名: | 曹洁 黄开杰 王进花 |
| |
作者单位: | 兰州理工大学计算机与通信学院,兰州730050;兰州理工大学电气工程与信息工程学院,兰州730050;兰州理工大学电气工程与信息工程学院,兰州,730050 |
| |
基金项目: | 国家自然科学基金资助项目(61633031,61763028);甘肃省自然科学基金资助项目(1506RJZA105) |
| |
摘 要: | 为了解决粒子滤波多说话人跟踪过程中粒子易发散导致多目标跟踪精度低的问题,提出了并行粒子滤波和基于GPU的K-均值聚类的多声源定位方法。该方法首先分析了粒子滤波在实现多目标跟踪时,进行数据关联的过程产生较大的计算量,并且出现多个目标时,粒子会逐渐发散。针对计算量大和粒子发散的问题,提出了一种并行粒子滤波和K-均值聚类的方法。实验表明,随着粒子数和目标数的增加,计算量以指数增加,并且粒子发散严重,采用基于GPU的K-均值聚类方法的粒子滤波多说话人跟踪方法,相比传统粒子滤波跟踪方法具有更收敛的粒子集并且跟踪精度较高。
|
关 键 词: | GPU 粒子滤波 K-均值 多目标跟踪 |
收稿时间: | 2017-10-24 |
修稿时间: | 2019-01-25 |
Particle filter multi-speakers tracking algorithm based on GPU and its application |
| |
Affiliation: | Lanzhou University of Technology |
| |
Abstract: | In order to solve the problem of low accuracy of multi-target tracking, particles in particle filter are easy to disperse in the process of multi-speaker tracking. This paper presented a parallel particle filter algorithm and GPU-based K-means clustering multi-source localization method. The method first analyzed the particle filter to achieve multi-target tracking, data association process have a large amount of computation, and the particles gradually diverge with the emergence of multiple targets. In order to solve the problem of large amount of computation and particle divergence, this paper proposed a method of parallel particle filter and k-means clustering. Experiments show that, as the number of particles and the number of targets increases, the amount of computation increases exponentially and the particles scatter seriously. In this paper, by using the GPU-based k-means clustering method, the particle filter multi-speaker tracking method has more convergent particle sets and higher tracking accuracy than the traditional particle filter tracking method. |
| |
Keywords: | GPU particle filter k-means multi-target tracking |
本文献已被 万方数据 等数据库收录! |
| 点击此处可从《计算机应用研究》浏览原始摘要信息 |
|
点击此处可从《计算机应用研究》下载全文 |
|