基于圆形拟合识别算法的颗粒粒度粒形自动测量方法研究  

Automatic Particle Measurement Method Based on Circle Fitting Aided Recognition Algorithm

作　　者：许潇[1] 张文阁[1] 池顺鑫 XU Xiao;ZHANG Wenge;CHI Shunxin(National Institute of Metrology,Beijing 100029,China)

机构地区：[1]中国计量科学研究院,北京100029

出　　处：《计量科学与技术》2021年第3期15-18,共4页Metrology Science and Technology

基　　金：国家重点研发计划(2016YYF0204304);国家自然科学基金青年基金资助项目(21505125)。

摘　　要：球形颗粒的粒度粒形通常采用显微成像和图像分析法进行测量,需由人工处理每个颗粒图像,操作耗时且会引入人为误差。提出了基于圆形拟合辅助识别算法的颗粒粒度粒形自动测量方法,采用圆形拟合预先识别显微图像中的单个颗粒,并以拟合结果为指导优化颗粒边缘识别,从而实现球形颗粒的粒度粒形自动测量。与人工处理结果相比,自动测量在确保测量结果一致性的同时极大提高了测量速度与重复性。The microscope imaging and image analysis method was commonly used for the size and geometry measurement of spherical particles.As manual processing for particle images were required,the measurement procedure was time-consuming,and extra uncertainties were introduced.To solve this problem,the automatic particle measurement method based on circle fitting aided recognition algorithm was developed.The automatic particle size and geometry analyses were realized by using circle fitting aided single-particle recognition,followed by the optimization of edge recognition.Comparing with the manual method,the automatic method greatly speeds up the image processing procedure with better repeatability.

关 键 词：颗粒测量 粒度粒型 图像处理 圆形拟合算法 颗粒边界识别 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]