基于图像处理的蔗糖结晶颗粒识别方法  

A Method to Identify Sucrose-Crystallizing Particles Based on Image-Processing

作　　者：张振升[1] 朱名日[1] 潘泽锴[1] 

机构地区：[1]桂林电子科技大学计算机与控制学院,广西桂林541004

出　　处：《计算机系统应用》2010年第3期95-99,34,共6页Computer Systems & Applications

基　　金：广西教育厅科研基金(200808LX138);广西科技厅科技攻关基金(0428007-9)

摘　　要：针对煮糖过程蔗糖结晶图像的特点采用颜色空间转换的方法,将图像从RGB颜色空间转换为HIS颜色空间,利用HIS颜色空间各分量相对独立性以及结晶颗粒和糖浆溶液的色调差异通过改进的大津法对H分量进行阈值分割,再通过数学形态学,中值滤波,孔洞填充,去除噪声颗粒等方法进行后续处理,最后对处理后的图像提取结晶颗粒的个数、面积、周长、形状指数以及面积比等特征值并进行分析来完成煮糖过程中蔗糖结晶颗粒的识别。实验证明该方法识别率高,切实可行。Based on the characteristics of sucrose-crystallizing images in the sugar-boiling process,this paper uses a method called color space conversion which changes the images from RGB color space into HIS color space,and makes use of the relative independence of every component in the HIS color space and the color difference between crystal particles and syrup solution to segment the H component with the improved Otsu＇s thresholding method.Many ways are used in the subsequent processing,such as mathematical morphology opening operation,median filter,holes filled,noise removing process of particles in the subsequent processing.Finally,the identification of sucrose-crystallizing particles in the process of boiling sugar is made by analysing the feature value such as the number,size,perimeter,shape index and the area ratio of the crystalline particles extracted from the processed images.Experiments show that this method is effective and practical.

关 键 词：HIS 改进的大津法 数学形态学 孔洞填充 噪声颗粒 特征提取 结晶颗粒识别 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] TQ133.1[自动化与计算机技术—计算机科学与技术]