| 基于深度学习的龙眼叶片叶绿素含量预测的高光谱反演模型 |
| |
| 引用本文： | 甘海明,岳学军,洪添胜,凌康杰,王林惠,岑振钊.基于深度学习的龙眼叶片叶绿素含量预测的高光谱反演模型[J].华南农业大学学报,2018,39(3):102-110. |
| |
| 作者姓名： | 甘海明 岳学军 洪添胜 凌康杰 王林惠 岑振钊 |
| |
| 作者单位： | 南方农业机械与装备关键技术教育部重点实验室;华南农业大学电子工程学院;华南农业大学工程学院 |
| |
| 基金项目： | 国家自然科学基金（30871450）；广东省科技计划项目（2015A020224036，2014A020208109）；广东省水利科技创新项目（2016-18） |
| |
| 摘 要： | 【目的】探讨龙眼Dimocarpus longan Lour.叶片发育过程中叶绿素含量二维分布变化规律,实现无损检测病虫害对叶片叶绿素含量分布的影响,为评估嫩叶抗寒能力、龙眼结果期的施肥量和老熟叶的修剪提供参考。【方法】利用高光谱成像仪采集龙眼叶片在369~988 nm区间的高光谱图像,自动提取感兴趣区域,利用分光光度法测定叶片叶绿素含量。基于皮尔森相关系数(r)分析了龙眼叶片生长过程中各波段光谱响应与叶绿素含量之间相关性,建立偏最小二乘回归模型。分析了特征波段图像纹理特征与叶绿素含量相关性,将光谱特征和纹理特征结合导入深度学习中的稀疏自编码(SAE)模型预测龙眼叶片叶绿素含量,结合"图谱信息"的SAE模型预测龙眼叶片叶绿素含量的分布情况。【结果】龙眼叶片3个生长发育期相关系数的曲线均在700 nm附近出现波峰,嫩叶、成熟叶和老熟叶3个阶段相关性最高的波长分别为692、698和705 nm;全发育期的最敏感波段相关性远高于3个生长发育期,r达到0.890 3。回归模型中,吸收带最小反射率位置和吸收带反射率总和建立的最小二乘回归模型预测效果最好(R_c~2=0.856 8,RMSEc=0.219 5;R_v~2=0.771 2,RMSEv=0.286 2),其校正集和验证集的决定系数均高于单一参数建立的预测模型。在所有预测模型中,结合"图谱信息"的SAE模型预测效果最好(R_c~2=0.979 6,RMSEc=0.171 2;R_v~2=0.911 2,RMSEv=0.211 5),且预测性能受叶片成熟度影响相对较小,3个生长阶段R_v~2的标准偏差仅为最小二乘回归模型标准偏差的29.9%。【结论】提出了一种自动提取感兴趣区域的方法,成功率为100%。基于光谱特征的回归模型对不同生长阶段的叶片预测效果变化较大,而基于"图谱信息"融合的SAE模型预测性能受叶片成熟度影响相对较小且预测精度较高,SAE模型适用于不同成熟度的龙眼叶片叶绿素含量分布预测。
|
| 关 键 词： | 龙眼叶片 高光谱成像 叶绿素含量 光谱特征 图像纹理特征 反演 |
| 收稿时间： | 2017/9/14 0:00:00 |
A hyperspectral inversion model for predicting chlorophyll content of Longan leaves based on deep learning |
| |
| GAN Haiming,YUE Xuejun,HONG Tiansheng,LING Kangjie,WANG Linhui and CEN Zhenzhao.A hyperspectral inversion model for predicting chlorophyll content of Longan leaves based on deep learning[J].Journal of South China Agricultural University,2018,39(3):102-110. |
| |
| Authors: | GAN Haiming YUE Xuejun HONG Tiansheng LING Kangjie WANG Linhui and CEN Zhenzhao |
| |
| Institution: | Key Laboratory of Key Technology on Agricultural Machine and Equipment, Ministry of Education, Guangzhou 510642, China;College of Electronic Engineering, South China Agricultural University, Guangzhou 510642, China,Key Laboratory of Key Technology on Agricultural Machine and Equipment, Ministry of Education, Guangzhou 510642, China;College of Electronic Engineering, South China Agricultural University, Guangzhou 510642, China,Key Laboratory of Key Technology on Agricultural Machine and Equipment, Ministry of Education, Guangzhou 510642, China;College of Engineering, South China Agricultural University, Guangzhou 510642, China,Key Laboratory of Key Technology on Agricultural Machine and Equipment, Ministry of Education, Guangzhou 510642, China;College of Electronic Engineering, South China Agricultural University, Guangzhou 510642, China,Key Laboratory of Key Technology on Agricultural Machine and Equipment, Ministry of Education, Guangzhou 510642, China;College of Electronic Engineering, South China Agricultural University, Guangzhou 510642, China and Key Laboratory of Key Technology on Agricultural Machine and Equipment, Ministry of Education, Guangzhou 510642, China;College of Electronic Engineering, South China Agricultural University, Guangzhou 510642, China |
| |
| Abstract: | |
| |
| Keywords: | Longan leaf hyperspectral image chlorophyll content spectral characteristic spectroscopy and texture feature inversion |
| 本文献已被 CNKI 等数据库收录！ |
| 点击此处可从《华南农业大学学报》浏览原始摘要信息 |
| 点击此处可从《华南农业大学学报》下载免费的PDF全文 |
|
