摘要
设计一种基于TMS320VC5416与FPGA的实时图像跟踪系统,该系统首先采用SAA7111A采集图像,经FPGA(采用乒乓操作的方式)传输给DSP处理器。在跟踪算法上采用灰度作为目标特征,使用粒子滤波方法实现对动态目标的跟踪。采用浮点数定点化处理,提高了算法的运行速度。实验表明:该系统具有良好的实时性与稳定性。
Based on TMS320C5416 and FPGA, a practical system for tracking of target using particle filter is presented. The original image is firstly sampled by SAA7111A, and then transmitted to DSP by Ping-Pang Method via FPGA. The gray feature is selected as the object characteristic and particle filter is used for object tracking. Fixed-point operation is used instead of floating-point operation for increasing the speed of computation. Experimental results show that the system operates well in real-time and is robust.
出处
《电视技术》
北大核心
2007年第3期85-87,共3页
Video Engineering
基金
国家自然科学基金资助(60372062)
关键词
实时跟踪
粒子滤波
模板匹配
real-time tracking
particle filter
template matching
