农田秸秆覆盖率检测方法与试验  

Method and test of straw coverage detection in farmland

作　　者：刘现悦 刘卉[1] 王侨 高娜娜[3] 梅鹤波 王培 Liu Xianyue;Liu Hui;Wang Qiao;Gao Nana;Mei Hebo;Wang Pei(College of Information Engineering,Capital Normal University,Beijing,100048,China;Beijing Research Center for Intelligent Agricultural Equipment,Beijing,100097,China;Beijing Research Center for Information Technology in Agriculture,Beijing,100097,China)

机构地区：[1]首都师范大学信息工程学院,北京市100048 [2]北京农业智能装备工程技术研究中心,北京市100097 [3]北京农业信息技术研究中心,北京市100097

出　　处：《中国农机化学报》2021年第2期203-208,共6页Journal of Chinese Agricultural Mechanization

基　　金：国家重点研发计划项目(2016YFD020060402);现代农业产业技术体系建设项目(CARS—02—87)。

摘　　要：针对秸秆覆盖率检测准确率易受光照不均匀影响以及现有检测方法对无秸秆图像检测精度较差的问题,提出一种基于图像行平均灰度标准差分类及图像分块可变阈值检测法。首先通过计算图像的行平均灰度标准差对图像进行分类,将标准差小于阈值的图像判断为无秸秆覆盖;然后将标准差大于阈值的图像划分成多个子块,对每个子块分别采用最大类间方差法进行阈值分割;最后将所有子块重新组合,用基本的形态学法计算目标图像的秸秆覆盖率。试验结果表明,该方法在很大程度上减少光照不均匀对检测结果的影响,对秸秆均匀分布图像的检测误差为2.87%,对无秸秆覆盖图像的检测误差为0.45%,能够提高农田秸秆覆盖率检测方法的适用性。In order to solve the problems that straw detection accuracy is easily affected by uneven illumination and that the existing detection methods have poor detection accuracy for straws without images.This paper proposes a method of image line average gray level standard deviation classification and image block variable threshold method.Firstly,the images were classified by calculating the average gray standard deviation of the rows of the images,and the image whose standard deviation is less than the threshold value is judged as no straw mulching.Then the image whose standard deviation is greater than the threshold is divided into several sub blocks,and each sub block is segmented by Otsu algorithm.Finally,all the sub-blocks are recombined and the straw coverage of the target image is calculated by the basic morphological method.The experimental results show that the method greatly reduces the influence of uneven illumination on the detection results.The detection error of uniform straw distribution image is 2.87%,and the detection error of no straw mulching image is 0.45%,which can improve the applicability of the detection method of crop straw coverage rate.

关 键 词：秸秆覆盖率 阈值分割 最大类间方差法 图像分块 保护性耕作 

分 类 号：S126[农业科学—农业基础科学] TP751[自动化与计算机技术—检测技术与自动化装置]