欢迎访问《中国农学通报》,

中国农学通报 ›› 2023, Vol. 39 ›› Issue (22): 152-157.doi: 10.11924/j.issn.1000-6850.casb2023-0212

• 农业信息·科技教育 • 上一篇    下一篇

基于OpenCV的‘翠冠’梨多角度大小检测与分级

刘现1(), 陈峻生2, 赵宇慧2   

  1. 1 福建省农业科学院数字农业研究所,福州 350003
    2 福建农林大学计算机与信息学院,福州 350003
  • 收稿日期:2023-03-15 修回日期:2023-05-22 出版日期:2023-08-05 发布日期:2023-07-28
  • 作者简介:

    刘现,女,1985年出生,福建福州人,助理研究员,硕士,研究方向:环境感知与智能控制。通信地址:350003 福建省福州市鼓楼区华林路188号 福建省农业科学院科技干部培训中心(数字农业研究所),Tel:0591-87869364,E-mail:

  • 基金资助:
    福建省农业科学院自由探索科技创新项目“基于大数据的翠冠梨智能分级模型构建”(ZYTS202234); 福建省农业科学院科技创新团队 “智慧农业科技创新团队”(CXTD2021013-1); 福建省农业科学院科技创新团队“南方丘陵农情监测科技创新团队”(CXTD2021012-3)

Multi-angle Volume Detection and Grading of ‘Cuiguan’ Pear Based on OpenCV

LIU Xian1(), CHEN Junsheng2, ZHAO Yuhui2   

  1. 1 Digital Agriculture Research Institute, Fujian Academy of Agricultural Sciences, Fuzhou 350003
    2 School of Computer Science and Technology, Fujian Agriculture and Forest University, Fuzhou 350003
  • Received:2023-03-15 Revised:2023-05-22 Online:2023-08-05 Published:2023-07-28

摘要:

为提高‘翠冠’梨大小检测以及分级的智能化程度与效率,基于OpenCV机器视觉技术综合多角度进行‘翠冠’梨大小检测与分级研究。在自主构建的图像采集系统试验平台上多角度采集100个‘翠冠’梨图像作为试验样本,基于图像使用Python语言结合OpenCV计算机视觉库编程求解‘翠冠’梨单角度果实大小像素值,综合多角度对果实大小像素值进行测算,参照‘翠冠’梨福建省地方标准对其进行分级并与传统的果实质量分级法进行对比来验证分级的效果。试验结果表明,基于‘翠冠’梨图像采用机器视觉技术综合多角度来进行果实大小检测与分级是一种具有一定可行性地智能化分级方法,达到了97%的分级正确率。研究结果可为基于机器视觉的翠冠梨大小分级方法提供技术参考。

关键词: ‘翠冠’梨, 图像, OpenCV, 分级

Abstract:

In order to improve the intelligence and efficiency of ‘Cuiguan’ pear size detection and grading, the research on ‘Cuiguan’ pear size detection and grading was conducted based on OpenCV machine vision technology from multiple perspectives. On the self-developed image acquisition system test platform, 100 images of ‘Cuiguan’ pear were collected from multiple angles as test samples. Based on the images, Python language and OpenCV computer vision library programming were used to solve the pixel values of the single angle size of ‘Cuiguan’ pear. Then the fruit size pixel values were calculated from multiple angles. To verify the effect of grading, ‘Cuiguan’ pear were graded according to the local standards of Fujian Province and compared with the traditional weight grading method. The test results show that using machine vision technology to detect and grade the fruit size from multiple angles based on the ‘Cuiguan’ pear image is an intelligent grading method with certain feasibility, reaching a grading accuracy of 97%. The research results can provide technical reference for the size grading method of ‘Cuiguan’ pear based on machine vision.

Key words: ‘Cuiguan’ pear, image, OpenCV, grading