摘要
在VC++6.0开发平台上使用OpenCV函数库中基于加权颜色直方图的粒子滤波算法,设计了一个视觉跟踪器,完成了对人体图像序列中目标的动态跟踪。目标颜色模型的合理建立使得视觉跟踪器对系统资源需求较少,计算速度较快,利于实现实时跟踪。实验结果验证了该视觉跟踪器的有效性和实时性,它能够比较精确地实现基于观测量和控制量的后验概率分布,可用于解决即时定位与地图构建(SLAM)问题。
In this paper, the VC ＋ ＋6.0 development platform using OpenCV library weighted color histogram-based particle filter algorithm to achieve a visual tracking, dynamic tracking of targets in the human image sequences. Reasonable establishment of the tar- get color model makes less demand on system resources, fast calculation, visual tracking conducive to the realization of real-time track- ing. The experimental results verify the effectiveness and real-time of the visual tracking and it is able to more accurately express the posterior probability distribution based on the concept of measurement and control. It can be used to solve the SLAM problem.
出处
《电子技术应用》
北大核心
2013年第11期142-144,共3页
Application of Electronic Technique
基金
国家自然科学基金资助项目(61272364)
深圳市科技计划项目及基础研究计划项目(JC201005270275A)
深圳市战略性新兴产业发展专项资金项目(JCYJ20120614144655154)
北京师范大学珠海分校科研创新团队项目(201251006)
北京师范大学珠海分校教改项目(201329)
