摘要
以猴头菇为研究对象,在计算机视觉技术和图像处理技术的基础上,对猴头菇进行无损检测和分级.采用灰度化、中值滤波、二值化、边缘提取等技术对图像进行预处理,提取了猴头菇面积、周长、颜色、径长、形状等特征参数.以猴头菇的颜色、大小、形状作为分级指标,确定并完善了猴头菇分级标准.根据猴头菇的分级标准,在Fisher判别分析基础上构建了猴头菇的分级模型.结果表明,基于计算机视觉技术的猴头菇无损检测与分级结果准确率均在80%以上.
Taking Hericium erinaceus as the research object,using technologies of computer vision and image process,nondestructive hericium erinaceus testing and classification were carried out.The images were preprocessed by grayscale,median filter,binarization,edge extraction and other techniques,and the characteristic parameters such as area,perimeter,color,diameter,and shape of Hericium erinaceus were extracted.Taking the color,size and shape of Hericium erinaceus as the grading index,hericium erinaceus classification standard was determined and perfected.According to the classification standard,hericium erinaceus classification model was constructed on the basis of Fisher discriminant analysis.The results show that the accuracy of non-destructive hericium erinaceus testing and classification based on computer vision technology is above 80%.
作者
王锦谟
孙涛
邓成
Wang Jinmo;Sun Tao;Deng Cheng(Department of Mechanical and Electrical Engineering,Anhui Vocational College of Grain Engineering,Hefei 230011,China)
出处
《洛阳师范学院学报》
2022年第8期21-25,共5页
Journal of Luoyang Normal University
基金
安徽省高等学校自然科学研究项目(KJ2020A1140)
安徽省高等学校自然科学研究项目(KJ2021A1559)。
关键词
猴头菇
计算机视觉
分级
预测模型
MATLAB
Hericium erinaceus
computer vision
classification
predictive model
MATLAB
