基于小波变换和分数阶微分的冬小麦叶绿素含量估算  

Estimation of Chlorophyll Content in Winter Wheat Based on Wavelet Transform and Fractional Differential

作　　者：李长春[1] 施锦锦 马春艳[1] 崔颖琪 王艺琳 李亚聪 LI Changchun;SHI Jinjin;MA Chunyan;CUI Yingqi;WANG Yilin;LI Yacong(School of Surveying and Land Information Engineering, Henan Polytechnic University, Jiaozuo 454000, China)

机构地区：[1]河南理工大学测绘与国土信息工程学院,焦作454000

出　　处：《农业机械学报》2021年第8期172-182,共11页Transactions of the Chinese Society for Agricultural Machinery

基　　金：国家自然科学基金项目(41871333);国家大学生创新创业项目(202010460048)。

摘　　要：叶绿素含量变化直接表征冬小麦的光合作用能力,所以监测冬小麦叶绿素含量对分析冬小麦光合能力和生长状况具有重要意义。基于地面冬小麦冠层高光谱和实测叶绿素含量,分别利用原始光谱、分数阶微分光谱、原始光谱经连续小波变换后得到的小波能量系数与实测叶绿素含量进行相关性分析,选取相关性较好的分数阶微分光谱和小波能量系数,采用逐步回归分析、支持向量机、人工神经网络等方法构建冬小麦叶绿素含量估算模型。结果表明,在拔节期、孕穗期、开花期和全生育期,使用连续小波变换人工神经网络建模结果最优,拔节期建模和验证决定系数分别为0.93和0.90,孕穗期建模和验证决定系数分别为0.93和0.90,开花期建模和验证决定系数分别为0.93和0.90,全生育期建模和验证决定系数分别为0.86和0.85;在灌浆期,使用分数阶微分人工神经网络建模结果最优,灌浆期建模和验证决定系数分别为0.97和0.90。本研究可为作物叶绿素含量遥感估算提供技术方案。Chlorophyll content is the main biochemical parameter of winter wheat,and its changes directly represent the photosynthetic capacity of winter wheat.Therefore,monitoring the chlorophyll content of winter wheat is of great significance for analyzing the photosynthetic capacity and growth status of winter wheat.Based on the canopy hyperspectral data and measured chlorophyll content of winter wheat on the ground,the correlation analysis between the measured chlorophyll content and the wavelet energy coefficient obtained from the original spectrum,fractional differential spectrum and original spectrum through continuous wavelet transform was carried out,and then the fractional differential spectrum and wavelet energy coefficient with good correlation were selected and combined with stepwise regression analysis the estimation model of chlorophyll content of winter wheat was established by using the methods of support vector machine and artificial neural network.The results showed that:at the jointing stage,booting stage,flowering stage and full growth stage,the results of continuous wavelet transformartificial neural network modeling were the best,R^(2) of modeling and verification were 0.93 and 0.90 at jointing stage,0.93 and 0.90 respectively at booting stage,and 0.93 and 0.90 respectively at flowering stage,0.86 and 0.85 at full growth stage respectively;at the filling stage,the results of fractional differentialartificial neural network were the best,R^(2) of modeling and verification were 0.97 and 0.90,respectively,which provided technical scheme for remote sensing estimation of crop chlorophyll content.

关 键 词：冬小麦 叶绿素含量 冠层高光谱 连续小波变换 分数阶微分 人工神经网络 

分 类 号：S512.11[农业科学—作物学]