摘要
国家标准规定玉米的质量定等指标为容重,为了实现利用机器视觉快速预测玉米等级,采用自行构建的基于机器视觉技术的玉米检测系统获取4种不同等级的玉米籽粒图像,通过均值滤波、最大类间方差法和形态学运算对玉米籽粒和背景进行处理、分割和特征参数的选取,并采用主成分分析法确定图像特征信息的最佳主成分因子数,建立以玉米容重为基础的8-21-4三层BP神经网络质量等级识别模型。结果表明:利用BP神经网络对基于完整籽粒图像和籽粒横切面图像的玉米等级的总体识别率均在90%以上,因此利用该模型对玉米等级的检测识别具有较高的可行性。
Volume weight is the main index of maize quality grades according to national standard. In or- der to discriminate the maize grades accurately by machine vision, the grain image of four different grades of maize were adopted by the method of image processing with the industrial camera, the kernels and their background were processed, divided by the average filter, Otsu and morphological operation, the charac- teristic parameters were selected. The number of optimal main factors was determined by principal com- ponent analysis (PCA). The 8 -21 -4 three layers BP neural network model was established for the i- dentification of maize grades based on volume weight. Results showed that the overall recognition rate was over 90% based on the image of complete kernel and the kernel transverse section by BP neural network. So the model had high feasibility for detecting maize grades.
作者
周鸿达
张玉荣
王伟宇
陈赛赛
ZHOU Hong - da ZHANG Yu - rong WANG Wei - yu ZHOU Xian - qing CHEN Sai - sai(College of Food Science and Technology, Engineering Research Center of Grain Storage and Security of Ministry of Education, Grain Storage and Logistics National Engineering Laboratory, Henan University of Technology, Zhengzhou Henan 450001)
出处
《粮油食品科技》
2016年第6期50-56,共7页
Science and Technology of Cereals,Oils and Foods
关键词
玉米
质量等级
机器视觉
主成分分析
BP神经网络
maize
quality grades
machine vision
principal component analysis ( PCA )
BP neural network