摘要
提出了适用于多目标的FPGA+DSP质心定位改进算法.该算法首先采用连通域标记算法在FPGA中标记出目标,解出各目标点的质心坐标,再由DSP根据统计结果进行自适应阈值计算.实验结果表明:多目标质心定位系统的检测结果与理论分析基本一致,可满足最高70fps的图像速率,该速率远高于基于单个DSP芯片或者普通计算机的处理速率.
In this paper,the improved FPGA and DSP centroid localization algorithm that is suitable for multiple target is proposed.The goals are marked out firstly in FPGA with the connected domain algorithm,and then the centroid coordinates of each target is worked out,after that,the coordinates adaptive threshold value is calculated according to the results of the statistics by DSP.The experimental results show that the multiple target centroid localization system test results are consistent with theoretical results that can reach the highest 70 fps rate of image,which is much higher than that on a single DSP chip or normal computer.
作者
周光宇
程萍
ZHOUGuang -yu;CHENG Ping(Ningbo Dahongying University,Ningbo 315175 China)
出处
《湘潭大学自然科学学报》
CAS
2018年第2期28-31,共4页
Natural Science Journal of Xiangtan University
基金
浙江省自然科学基金项目(LQ14F040004)
关键词
质心定位
直方图统计
自适应阈值
连通域标记
centroid localization
histogram statistics
adaptive threshold
connected component labeling