基于外形几何特征的鸡蛋重量检测模型  

Study on the Model of Egg Weight Detecting Based on its Geometry

作　　者：饶秀勤[1] 岑益科[1] 应义斌[1] 

机构地区：[1]浙江大学生物系统工程与食品科学学院,浙江杭州310029

出　　处：《中国家禽》2007年第5期18-20,共3页China Poultry

基　　金：浙江省科技计划项目(项目编号:2006C21050)

摘　　要：建立以外形几何特征为基础的鸡蛋重量检测模型。采集了鸡蛋图像,采用HIS颜色模型的H分量作为鸡蛋图像的阈值分割指标,用拉普拉斯算子提取出鸡蛋边缘,从鸡蛋边缘分别提取纵径、横径、上横径和下横径等4个鸡蛋外形几何特征量,用于建立鸡蛋重量检测的回归模型。试验表明,鸡蛋重量检测的绝对误差在±3g以内,能满足鸡蛋分级生产要求,可用于鸡蛋分级生产。A model for egg weight detecting based on its geometry was setup. An Egg image was grabbed,and then was segmented using H value threshold,and the edge was detected with Laplace operator. The vertical diameter,horizontal diameter,upper horizontal diameter and nether horizontal diameter was detected to setup a model between these value and egg weight. A test was done and validated the model. The absolute error was less than 3 g. The model could be used for egg grading.

关 键 词：鸡蛋 机器视觉 重量检测 

分 类 号：S83[农业科学—畜牧学]