基于机器视觉的哈密瓜大小分级研究

Hami melon Size Classification Based on Machine Vision

采用机器视觉技术对新疆哈密瓜进行自动大小分级。线阵相机在线采集哈密瓜样本RGB图像,通过对哈密瓜RGB图像进行灰度化、中值滤波、二值化、去除果梗、特征提取等一系列处理,获得哈密瓜二值化图像。利用椭圆拟合算法对二值化图像进行椭圆特征提取,基于椭圆长轴和椭圆率建立了哈密瓜大小分级标准,并以固定阈值建立分级模型。通过哈密瓜分级机系统进行大小分级,分级准确率达90.29%。

The automatic size grading of Xinjiang Hami Melon by using the machine vision technology.Line array camera online acquisition of Hami melon sample RGB image through gray,median filter,binary values,removal of fruit stem,feature extraction,and so on a series of processing of Hami melon RGB image,Hami melon binary image is obtained.Elliptic eigenvalue extraction of binary image,to the major axis of the ellipse and elliptic rate of Hami melon geometry size grading standard was established based on and with a fixed threshold established classification model based on ellipse fitting algorithm.The geometric size grading by Hami melon classification system,classification accuracy reached 90.29%.