棉花产量遥感预测的L-Y模型构建

本文利用LAI动态与棉花产量的关系建立了叶面积指数—产量（L-Y）模型，以期利用多时相遥感数据，实现对棉花产量定量遥感预测。模型建立以小区控制和大田生产试验数据为基础，以农学原理为背景，采用数学推演方法，具简单、灵活、普适性强等特点。检验结果表明，用便携式光谱辐射计测定棉花冠层高光谱反射率，以棉花全生育期LAI动态与棉花产量的关系和近地高光谱遥感参数模型监测的多时相LAI，可很好地定量预测棉花产量，估算误差约为5.44%，RMSE达到116.2 kg·hm^-2，预测值与实测值相关系数为0.836，达极显著水平。L-Y模型为棉花卫星遥感估产提供了参考模型，对其他作物使用动态生长信息提高遥感估产水平也有一定的借鉴意义。

L-Y Model of Cotton Yield Estimation by Remote Sensing

The model of L-Y according to the relationship between the dynamic leaf area index and the cotton yield was established based on the experiment data of the plot and field production, and the agricultural theory. The model has the characteristics of simplicity, flexibility, agility, and so forth. The results showed that the model could accomplish the qualitative estimation in the cotton yield with the measurement of leaf area index, the cotton yield and the multi-time LAI from the cotton canopy hyperspectral reflectance by field spectrum radiometer. The relative error of estimation was about 5.44%, and RMSE was 116.2 kg·ha^-1, and correlation coefficient between measured and estimated values was 0.836 with a high significance at the 0.99 confidence interval. The referenced model for yield estimation provided from the present research could contribute the improvement of other crop yield remote sensing estimation using dynamic growth information.