液滴图像处理与特征提取  

Liguid droplet image processing and feature extraction

作　　者：季宏兵[1] 扈映茹[2,3] 

机构地区：[1]上海交通大学电子信息与电气工程学院,上海201201 [2]成都天马微电子有限公司,四川成都611730 [3]陕西科技大学电气与信息工程学院,陕西西安710021

出　　处：《光学仪器》2010年第2期34-38,共5页Optical Instruments

基　　金：西安市科技局快速响应计划资助项目(ZX06084)

摘　　要：液体的物理化学性质直接影响到其液滴生长的变化,采用自主搭建的液滴图像检测系统监测液滴生长过程中的轮廓变化情况,并对液滴图像进行预处理和特征提取。用主成分分析(PCA)法对提取的面积、周长、液滴长度、宽度、矩形度、伸长度和圆形度等液滴轮廓特征参数进行了分析,得到了液滴长度、宽度、周长和面积四个有效特征参数。结果表明该方法用于液体分析和鉴别是可行性的。The physical and chemical properties of liquid droplet directly affect their growth. Droplet image detection system, which is built on our own, was used to monitor the surface profile change during droplet growth. Preprocessing and characterization of droplet image were done and characteristic parameters of surface profile such as area, perimeter, length, width, rectangularity, elongation and circularity were analyzed by the principal component analysis （PCA） method. The obtained four effective parameters such as length, width, perimeter, and area demonstrated the feasibility of PCA method in analysis and identification.

关 键 词：液滴分析 CCD 图像处理 特征提取 主成分分析(PCA) 

分 类 号：TH741.4[机械工程—光学工程]