基于双核机制的实时图像跟踪系统

摘要

为解决Camshift(continuously adaptive meanshift)跟踪算法在嵌入式平台上速度非常有限,很难实现实时性的问题,研究并实现基于双核的实时图像跟踪系统.使用双内核的DM3730为核心,实现DSP核与ARM核的协同工作,利用DSP强大的数值运算能力与双核的并行工作提高图像跟踪算法Camshift的实时性能.实验结果表明,该系统具有较强的抗干扰能力与较强的目标跟踪能力,在输入为320*240视频帧时,其平均处理能力达到了21.6 fps.%To solve the problem that Camshift is difficult to be realized in real time because of the limited operation speed on the embedded platform, the real-time image tracking system based on the dual core was researched and realized.Using dual kernel DM3730 as the core, this system realized the cooperative work between the DSP core and ARM core.The DSP's powerful numerical computation and the parallel work of dual core were used to improve the real-time performance of image tracking algorithm based on Camshift.Experimental results show that the image tracking system based on dual core has strong anti-interfe-rence ability and strong target tracking ability.The operating speed reaches as fast as 21.6 fps when the input of the video frame is 320*240.