无人机光谱分析在水稻产量预测中的应用

The Application of Unmanned Aerial Vehicle Spectral Analysis in Production Prediction of Rice

为了实现大范围水稻产量预测,采用无人机检测水稻冠层光谱,通过分析光谱建立水稻产量模型。叶绿素含量也可有效表征水稻生长情况,预测水稻产量,现采用无人机检测水稻冠层叶绿素光谱,发现在红边区域和近红外波段产生特征吸收波长。综合考虑上述特征吸收波长,采用差值植被指数作为建模因子,建立差值植被因子在水稻不同生长时期的线性相关系数云图,进行线性分析,发现差值植被指数在拔节期与抽穗期存在严重线性相关。因此,以孕穗期与乳熟期差值植被指数与水稻亩产进行逐次线性拟合,采用'留一法'对8组样本轮流预测,预测精度较高。对模型进行线性拟合评价和统计学评价,结果表明该模型可靠性高。

In order to achieve automatic monitoring for large range rice,rice canopy spectral was test and production prediction of rice was built by spectral analysis.The chlorophyll content can characterize the growth of rice and predict the rice yield.The rice canopy chlorophyll spectral was test by unmanned aerial vehicle,characteristic absorption wavelength distributed in red edge and near-infrared waveband.Difference vegetation index was used as factor of model.Cloud chart of linear correlation coefficient was achieved.Difference vegetation index in jointing stage and heading stage had great linear correlation by collinearity analysis.The Rice yield and difference vegetation index in booting stage and milk stage were taken to progressive regression analyses.8 samples were predicted by leaving-one method,and prediction accuracy was good.Linear fitting evaluation and statistical evaluation were used to test this model,the result showed this model had high reliability.