基于机器视觉的产地脏污鸭蛋外形扁平度在线检测  

Online Detecting and Grading Method of Shape Flatness of Dirty Duck Eggs Based on Machine Vision

作　　者：王巧华[1,2] 芦茜[1] 马美湖[2,3] 段宇飞[1] 王彩云[1] 

机构地区：[1]华中农业大学工学院,武汉430070 [2]国家蛋品加工技术研发分中心华中农业大学,武汉430070 [3]华中农业大学食品科学技术学院,武汉430070

出　　处：《中国食品学报》2017年第5期200-207,共8页Journal of Chinese Institute Of Food Science and Technology

基　　金：国家自然科学基金项目(31371771);公益性行业(农业)科研专项(201303084);国家科技支撑计划项目(2015BAD19B05)

摘　　要：产地鸭蛋在养殖地收集后需就地粗选,根据不同品质分类装箱运往蛋企加工厂清洗、包装及流通。产地鸭蛋表面脏污严重,粗选难度较大。本文采用机器视觉技术采集群体产地鸭蛋动态图像,利用R-B分量有效去除因装置间隙产生的漏光;对运行中的每枚蛋拍照3次,取最大面积图像,减小脏污位于图像边缘时对其边缘检测的影响;用最小平均中值法对产地鸭蛋边缘进行椭圆拟合,避免因光源过亮和鸭蛋过脏带来的边缘噪声的干扰;提取产地鸭蛋长半轴、短半轴、体积和离心率作为特征参数,通过LIBSVM建立产地鸭蛋外形扁平度分级模型,实现对产地鸭蛋尺寸和扁平程度的分级。经检验,鸭蛋尺寸与扁平度的分级正确率分别为91.67%,95.00%。Grading eggs by their shapes is an important part in processing chain. Duck eggs should be graded according to their shape in farm in order to pack them in suitable egg cartons, and then be transported to processing plants to clean, pack and put into market. Duck eggs usually stick with dirt, weeds and feather due to the habitat of duck which is around water. It is difficult to grade duck eggs for complex smudges on the surface. In order to solve this problem, an online detecting and grading method of shape of dirty duck eggs based on machine vision was developed. Light leak causing by the space of device was eliminated efficiently by deducting component B from component R. Every duck egg was photographed three times on conveyor, and the image with the largest area was chosen to be representa- tion. Ellipse fitting was conducted on the edge of dirty duck egg in each representative image through least median of squares algorithm. Through these image proeessings, undesirable effect caused by complex smudges on the surface of duck eggs was minimized. The values of semi-major axis, semi-minor axis, volume and eccentricity of fitted ellipse were extracted as characteristic parameters of dirty duck eggs. The model of dirty duck eggs was established by LIBSVM, which realized the grading of size and flat degree of dirty duck eggs. The correct grading rate was 91.65% and 95% respectively.

关 键 词：机器视觉 脏污鸭蛋 外形 边缘检测 分级 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] TS253.7[自动化与计算机技术—计算机科学与技术]