欢迎访问《农学学报》,

农学学报 ›› 2022, Vol. 12 ›› Issue (2): 73-75.doi: 10.11923/j.issn.2095-4050.cjas2020-0218

所属专题: 生物技术 智慧农业 数字乡村

• 农业信息/农业气象 • 上一篇    下一篇

数字图像处理技术在叶面积测量中的应用

宋英博()   

  1. 黑龙江省农业科学院佳木斯分院,黑龙江佳木斯 154007
  • 收稿日期:2020-11-26 出版日期:2022-02-20 发布日期:2022-03-16
  • 作者简介:宋英博,男,1979年出生,黑龙江双鸭山人,助理研究员,硕士,主要从事农业信息化与植物营养研究。通信地址:154007 黑龙江省佳木斯市东风区安庆街531号,E-mail: 1005768095@qq.com
  • 基金资助:
    黑龙江省农业科学院院级科研项目“欧洲硬粒型玉米种质评价及创新利用”(2019YYYF015);黑龙江省农业科学院院级科研项目“三江平原玉米最优施肥模型的研究与应用”(2020YYYF049)

Leaf Area Measurement System Based on Digital Image Processing Technology

SONG Yingbo()   

  1. Jiamusi Branch of Heilongjiang Academy of Agricultural Sciences, Jiamusi 154007, Heilongjiang, China
  • Received:2020-11-26 Online:2022-02-20 Published:2022-03-16

摘要:

传统测量叶面积方法费时、低效,叶面积仪法高成本、维修不便。本研究利用图像处理技术测量叶面积,从解决图像阈值的分割、叶片阴影去除以及叶片边缘检测算法等问题出发,应用大津法求得阈值,中值滤波法去除杂点,采用Roberts算子检测边缘,进而计算叶面积。叶面积仪法与图像处理法比较叶面积值相关系数R2为0.962,剪纸法与图像处理法比较叶面积值相关系数R2为0.949,最后验证本方法适合大量叶面积的测量工作,且具有速度快、数据准确、精度高的特点。

关键词: 图像处理, 阈值, 算法, 像素, 叶片

Abstract:

In view of the time consuming and low efficiency of the traditional method of leaf area measurement, and the high cost and inconvenient maintenance of the leaf area meter method, this study used the image processing technology to measure the leaf area. The computer vision technology reference method was used to solve the limitations of image threshold segmentation, leaf shadow removal and leaf edge detection algorithm, and the threshold was obtained by OTSU method, the median filtering method was used to remove residual impurities in the image, and the Roberts operator was used for edge extraction. Leaf areas are calculated and compared with those obtained by leaf area meter and paper-cutting method. The leaf area determined by the leaf area meter method and the image processing method had the correlation coefficient R2 of 0.962, and the leaf area determined by the paper cutting method and the image processing method had the correlation coefficient R2 of 0.949. The digital image processing method is validated to be suitable for large amount of leaf area measurement, and has fast speed, accurate data and high precision.

Key words: image processing, threshold, arithmetic, pixel, leaf