基于计算机视觉的铁矿石粉微粒尺寸的检测  

The measurement of ironstone particulates based on computer vision

作　　者：张学军[1] 陈向伟[1] 陈国军[1] 

机构地区：[1]东北电力大学能源与机械工程学院,吉林吉林132012

出　　处：《矿山机械》2009年第20期14-16,共3页Mining & Processing Equipment

基　　金：国家自然基金(10572053);东北电力大学博士基金(BSJXM-200714)

摘　　要：针对铁矿石粉微粒尺寸在几微米到几十微米的特点,提出了基于重力场流分离结合图像处理技术检测微粒尺寸的方法。首先利用重力场流分离系统分离铁矿石粉微粒,然后,利用光学生物显微镜与CCD采集铁矿石粉微粒图像,最后,利用图像处理软件对微粒图像进行预处理,用阈值化方法分割图像。利用动态聚类,使用C++语言自编程计算,对预处理后的图像作进一步处理,得出微粒个数、最大半径、最小半径、平均半径及半径方差等。结果表明,检测微粒尺寸参数的方法是方便可行的。The measurement for micro particulates is proposed based on image processing and gravitation field flow fractionation, aiming at the characteristics of ironstone particulates which sizes are from several microns to several tens microns. Firstly; the ironstone particulates are separated by gravitation fieldflowfractionation. Secondly, the images of the particulates are collected by a microscope and CCD, and then the particulates images are pretreated by image process software and segmented by threshold. At last, utilizing the dynamic clustering method coded by C＋＋, the results are obtained including the amounts of particulates, the maximum radius, minimum radius, average radius, standard deviation of radius, and so on. The results indicate that the presented method is correct and feasible.

关 键 词：微粒半径 场流分离 CCD 动态聚类 

分 类 号：TF041[冶金工程—冶金物理化学]