基于小波能量系数和叶面积指数的冬小麦生物量估算  

Winter Wheat Biomass Estimation Based on Wavelet Energy Coefficient and Leaf Area Index

作　　者：李长春[1] 李亚聪 王艺琳 马春艳[1] 陈伟男 丁凡 LI Changchun;LI Yacong;WANG Yilin;MA Chunyan;CHEN Weinan;DING Fan(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12期191-200,共10页Transactions of the Chinese Society for Agricultural Machinery

基　　金：国家自然科学基金项目(41871333);河南省科技攻关项目(212102110238)。

摘　　要：生物量是评价作物长势及产量估算的重要指标,科学、快速、准确地获取生物量信息,对于监测冬小麦生长状况以及产量预测等具有重要意义。以冬小麦为研究对象,通过相关性分析,选取相关性较好的小波能量系数,同时耦合叶面积指数,基于支持向量回归算法、随机森林算法、高斯过程回归3种算法构建冬小麦生物量估算模型。结果显示,基于小波能量系数,分别利用支持向量回归算法、随机森林算法、高斯过程回归进行生物量估算,4个生育期的验证R^(2)分别是0.55、0.40、0.39;0.75、0.70、0.83;0.84、0.92、0.93;0.84、0.89、0.85。表明高斯过程回归模型估算精度最优。叶面积指数耦合小波能量系数,利用支持向量回归算法、随机森林回归算法、高斯过程回归进行生物量估算,4个生育期的验证R^(2)分别是0.76、0.73、0.77;0.76、0.72、0.84;0.87、0.94、0.94;0.85、0.90、0.91。表明高斯过程回归算法估算精度最优,并且在一定程度上能够克服冠层光谱饱和现象,提高模型估算精度。以小波能量系数和叶面积指数为输入变量结合高斯过程回归算法建立冬小麦生物量估算模型,可以提高生物量估算精度,为基于遥感技术的作物参数快速估算提供参考。Biomass is an important indicator for evaluating crop growth and yield estimation.Obtaining biomass information scientifically,quickly and accurately is of great significance for monitoring the growth status of winter wheat and yield prediction.Taking winter wheat as the research object,through correlation analysis,the wavelet energy coefficient with good correlation was selected,and the leaf area index was coupled at the same time.Based on the support vector regression algorithm,random forest algorithm,and Gaussian process regression,three algorithms were used to construct a winter wheat biomass estimation model.The verification R^(2)of the four growth periods were 0.55,0.40 and 0.39;0.75,0.70 and 0.83;0.84,0.92 and 0.93;0.84,0.89 and 0.85,respectively.It was showed that the estimation accuracy of Gaussian process regression model was the best.Leaf area index coupled with wavelet energy coefficients,using the three algorithms to estimate biomass,the verification R^(2)of the four growth periods were 0.76,0.73 and 0.77;0.76,0.72 and 0.84;0.87,0.94 and 0.94;0.85,0.90 and 0.91,respectively,indicating that the Gaussian process regression algorithm had the best estimation accuracy,and to a certain extent,it can overcome the canopy spectrum saturation phenomenon and improve the estimation accuracy of the model.Using wavelet energy coefficient and leaf area index as input variables combined with Gaussian process regression algorithm to establish a winter wheat biomass estimation model,which can improve the accuracy of biomass estimation and provide a scientific reference for the rapid estimation of crop parameters based on remote sensing technology.

关 键 词：冬小麦 生物量 小波能量系数 叶面积指数 高斯过程回归 

分 类 号：S532[农业科学—作物学]