江西农业学报2024，Vol.36Issue(1)：35-41,7.DOI:10.19386/j.cnki.jxnyxb.2024.01.006
基于数字图像处理技术的叶片结构模型构建
Construction of Blade Structure Model Based on Digital Image Processing Technology
摘要
Abstract
In order to study the use of image method instead of traditional four-layer vibrating screen method to establish a leaf structure feature area model,the image of each unfolded tobacco leaf was scanned offline,the B-component index threshold and the corrosion expansion closure operation were carried out,and the complete tobacco leaf image was extracted by eliminating the background.The pixel area of each tobacco leaf was calculated and sorted by 8-neighborhood extraction.One point at every interval traversed all points in the pixel area interval,and compared with the actual large,medium,small and fragment rates.The pixel area value with the smallest prediction error under different area thresholds was selected as the area feature threshold model composed of large,medium,small and fragment rates,and external verification and repeatability detection were carried out on the model.The results show that the blade structure is composed of[+∞,50880],(50880,17230],(17230,2380],(2380,0]area characteristic interval model.Through external verification and comparison with the standard four-layer vibrating screen,The errors of large,medium,small and fragment rates were 1.09%,0.64%,0.47%and 1.07%,respectively.The standard deviations of large,medium,small and fragment rates were 0.88%,0.84%,0.19%and 0.26%,respectively.The blade structure model established by image method has good stability and can predict the large,medium,small and fragment rates of tobacco leaves well.关键词
图像/B分量/像素面积/叶片结构/振筛Key words
Image/B component/Pixel area/Blade structure/Vibrating screen分类
农业科技引用本文复制引用
胡春华,王发勇,张智强,罗姗,蔡永珍,吴贵轩,罗军,李一辉,詹映,刘晨凯..基于数字图像处理技术的叶片结构模型构建[J].江西农业学报,2024,36(1):35-41,7.基金项目
红云红河烟草(集团)有限责任公司科技项目"近红外与MSPC在打叶复烤均质化过程质量管控中的应用研究"(HYHH2021GY09). （集团）
