摘要
以大豆叶片叶绿素为对象,基于RGB色彩空间构建叶绿素SPAD估算模型,对大豆叶片叶绿素含量进行预测。首先,采集自然环境下的大豆叶片图像,运用中值滤波法去除图像噪声,并基于k-means算法将叶片从背景中分割;其次,提取叶片图像的红(R)、绿(G)、蓝(B)值,通过运算组合构建颜色特征参数,建立基于大豆叶片颜色特征参数的叶绿素含量估算模型,并对其精度进行评价和验证;最后,组合了R/G/B、R/(R+G-B)、B/(R+G-B)、G/(R+G-B)等4种颜色特征参数,利用这4种颜色特征参数和叶绿素实测值进行回归分析。结果表明,基于B/(R+G-B)的组合精确度最高,R2=0.438,AARD=9.58%,RMSE=2.862。该方法可以快速、无损地预测大豆叶片叶绿素含量,为评估大豆的生理状况提供了参考。
Taking the soybean leaves chlorophyll as the research object,the chlorophyll SPAD estimation model was constructed based on RGB color space.In this experiment,the prediction for chlorophyll content in soybean leaves was carried out.Firstly,the image of soybean leaves in the natural environment was collected,the image noise was removed by median filtering,and the leaves were segmented from the background based on the kmeans algorithm.Then,the red(R),green(G),and blue( B)values were extracted from the leaf images,the color feature parameters were constructed by operation combination,the chlorophyll content estimation model based on the color characteristic parameters of soybean leaves was established,and the accuracy was evaluated and verified.Finally,four color characteristic parameters such as R/G/B,R/(R+G-B),B/(R+G-B)and G/(R+GB)were combined, and the regression analysis was carried out on these four color characteristic parameters and chlorophyll measured values.The verification results showed that the RGB color space had the highest combination accuracy,R2 was 0.438,AARD was9.58%,and RMSE was 2.862.The method can predict the chlorophyll content in soybean leaves quickly and non-destructively,and it provides a scientific basis for assessing the physiological state of soybean.
作者
宋一帆
张武
姚雨晴
洪迅
张嫚嫚
刘连忠
SONG Yifan;ZHANG Wu;YAO Yuqing;HONG Xun;ZHANG Manman;LIU Lianzhong(School of Information and Computer,Anhui Agricultural University,Hefei 230036,Anhui,China)
出处
《江汉大学学报（自然科学版）》
2020年第1期65-72,共8页
Journal of Jianghan University：Natural Science Edition
基金
2018年安徽省重点研究和开发项目(1804a07020108)
2017年安徽省重大科技专项计划资助项目(17030701049)
2016年农业部农业物联网技术集成与应用重点实验室开放基金资助项目(2016KL05)
2019年安徽省重点研发计划面上攻关项目(201904a06020056)
