论文部分内容阅读
本研究的目的是通过数据同化方法,将国产环境卫星HJ-1A/B数据提取的水稻叶面积指数LAI信息和作物生长模型相结合,以提高水稻估产的精度.具体方法为:首先通过分析水稻叶面积指数LAI和水稻归一化植被指数NDVI的时域变化关系建立模型反演水稻LAI,并应用研究区历史数据和作物生长模型WOFOST建立初始水稻生长模型估算水稻产量.在构建代价函数的基础上,采用SCE(shuffledcomplex evolution)数据同化方法对初始水稻生长模型参数进行优化,使水稻生长模型估算的水稻LAI和遥感数据反演的LAI差值最小.最后将采用同化方法的水稻生长模型估算的研究区水稻产量和不加同化方法的原始水稻生长模型估算的水稻产量进行比较,结果显示水稻估产精度有明显提高.研究结果表明采用遥感数据提取的农作物实时生长信息可以修正作物生长模型关键参数以提高区域范围的农作物估产精度,同时也显示国产环境卫星数据在农作物生长监测上具有广阔的应用潜力.
The purpose of this study is to combine LAI information and crop growth model extracted from domestic environmental satellite HJ-1A / B data by data assimilation method in order to improve the precision of rice yield estimation.The specific method is as follows: Leaf area index (LAI) and normalized NDVI of rice (Oryza sativa L.) to establish a model to retrieve rice LAI and establish the initial rice growth model using the historical data and crop growth model WOFOST to estimate the rice yield.On the basis of constructing the cost function , The parameters of initial rice growth model were optimized by using data shuffled complex evolution (SCE) data to minimize the difference of LAI estimated by rice LAI and remote sensing data.At last, the rice growth model estimated by assimilation method Compared with the rice yield estimated from the original rice growth model without the assimilation method, the results showed that the rice yield estimation accuracy was significantly improved.The results showed that using the real-time crop growth information extracted from remote sensing data can correct the key parameters of crop growth model Increase crop yields on a regional scale Accuracy, but also shows that domestic environmental satellite data have a great potential in crop growth monitoring.