河南省冬小麦产量遥感监测精度比较研究

Comparative study on accuracy of winter wheat production by remote sensing monitoring in Henan province

关键词:  MODIS  实时植被指数  差值植被指数  冬小麦估产

Key words:  MODIS  real time NDVI  difference NDVI  estimation of winter wheat yield

摘要:

【目的】预测冬小麦产量，提高估产精度，为国家粮食生产安全宏观调控提供重要的 数据和技术支撑。【方法】以河南省为研究区，利用时间序列MODIS-NDVI 数据，结合河南 省17 个地市的产量数据以及河南省的冬小麦遥感监测面积数据，运用实时NDVI 与产量数 据建立线性模型、两年差值NDVI 与对应的差值产量建立线性模型两种方法对河南省冬小麦 产量进行预估，比较分析两种方法的精度。【结果】（1）两年差值归一化植被指数（dNDVI） 明显比实时归一化植被指数（NDVI）更好地预估冬小麦产量。（2）实时NDVI 预估单产产 量相对误差基本可以满足要求，大部分在0.25 以下，只有少数几个市（焦作、鹤壁、新乡、 安阳）相对误差较高。（3）两年差值归一化植被指数（dNDVI）预估单产的整体误差均可以 满足估产的要求，均在0.1 以下。【结论】两年差值归一化植被指数（dNDVI）可以有效预测 冬小麦产量，有利于提高估产精度。

Abstract:

［Purpose］Winter wheat is one of the Chinese main crops. The forecast of winter wheat yield and improvement of estimation accuracy are of great significance to the national macro-control of food production safety.［ Method］Based on the time series of MODIS-NDVI （Normalized Difference Vegetation Index） data combining with the production data of 17 cities in Henan province and remote sensing monitoring area data of winter wheat in Henan province，this paper takes Henan Province as the study area and constructs two methods to predict the winter wheat yield. One is to used real time-series NDVI and the corresponding yield to establish the linear model. The other is to take the difference NDVI between two years（ dNDVI） as model input variety and establish the linear winter wheat yield model with the corresponding yield. Then we compared the precision of the two methods and analysis model accuracy.［ Result］The results show：1） the yield model on the basis of taking the dNDVI as input variety was significantly better than that based on real time NDVI in predicting the yield of winter wheat；2） The relative error of the yield prediction model per unit can be satisfied by the real time NDVI，most of which were below 0.25，but only few cities（ i.e. Jiaozuo，Hebi，Xinxiang，and Anyang） have higher relative errors；3）The accuracy of winter wheat yield prediction model based on the dNDVI can meet the requirements of yield，these were below 0.1.［ Conclusion］Thus，the dNDVI can effectively predict the winter wheat yield and improve the yield estimation accuracy.