中文摘要：

多粒子追踪（MPT）是测量粒子布朗运动的一种方法。在200nm粒子扩散系数实验测量中,对MPT方法中的粒子识别和粒子匹配两个重要环节提出改进：在粒子识别时,通过调节灰度阈值并检测光斑面积的改变,可有效分别真实粒子和光晕引起的＂伪粒子＂;在粒子匹配中,将手动追踪和程序追踪得到的粒子单步位移用矢量法表示,可直观显示不同追踪方法的粒子匹配效果。实验结果表明,在MPT方法中增加上述措施后,手动追踪和程序追踪获得的实验扩散系数之间的偏差由8%~13%减小为2%~8%,说明基于阈值敏感性的＂伪粒子＂剔除方法合理、有效。

英文摘要：

Multiple-Particle Tracking（MPT） is a method for measuring particle Brownian motion.Based on the measurement of diffusion coefficients of 200nm particles,this paper presents two improvements in spot extraction and particle linking processes.For the spot extraction,by measuring the spot＇s area varieties when altering the gray-scale thresholds,the physical spots can be distinguished clearly from the ＂spurious spots＂ resulting from diffraction.For the particle linking,a single displacement of a particle can be presented as a vector.In this way the linked particle trajectories by MPT program and by hand can be compared obviously.The experiments show that the deviation of the diffusion coefficient obtained by MPT program from that by hand can be decreased from 8%～13% to 2%～8%,which approves the validity of the spot extraction method based on the gray-scale threshold sensitivity.