摘要
针对运动目标进行跟踪,常采用粒子滤波跟踪算法。为了减少相似背景像素点及光照变化对机动目标跟踪的干扰,采用了基于目标纹理特征和颜色特征融合的自适应粒子滤波算法,通过采用了不同的状态转移模型和观测模型,根据不同的跟踪环境进行自适应选择,并给出了相应的实验结果。实验结果表明,该算法在跟踪的性能和鲁棒性方面有所改进。
particle filter tracking algorithm is generally Used in tracking moving targets. In order to reduce the effect of similar background pixels and the light changes on maneuvering target tracking, the article uses adaptive particle filtering algorithm based on a target texture features, which is mixed with color features. By the help of different status transition model and observation model, this algorithm can select adaptively track target depending on the different track environment, and gives corresponding results. Test result shows that the algorithm in tracking performance and robu＇stness has great improvement.
出处
《智能计算机与应用》
2013年第2期54-57,共4页
Intelligent Computer and Applications
基金
东莞市科技计划项目(2012108102030)
关键词
目标跟踪
自适应选择
粒子滤波
直方图
鲁棒性
Target Tracking
Self Adaptive Selection
Particle Filtering
Histogram, Robustness
