基于GF-1影像的水稻拔节期地上部生物量估算

Estimation Biomass at Jointing Stage of Rice Based on GF-1 Images

目的:建立基于GF-1号WFV卫星影像数据的拔节期水稻生物量监测模型,为利用GF-1号影像进行拔节期水稻水肥管理以及生产力预测提供依据;方法:以安徽盛农农场水稻种植区域为研究对象,分析GF-1号WFV数据的光谱参数与地上部生物量直接的相关关系,筛选出较为精确的光谱参数,建立并评价水稻拔节期地上部生物量监测模型;结果:GF-1 WFV数据的光谱参数与地上部生物量具有良好的相关性。其中,直接利用单一波段建立的相关关系中,近红外的表现最好(R2=0.61),通过逐步多重线性回归分析发现利用绿光波段和近红外波段组合的函数能有效提高模型的精度(R2=0.659)。利用10种常见的植被指数与地上部生物量建立相关关系,发现NDVI、GNDVI、RVI、SAVI、EVI以及OSAVI与地上部生物量的相关性较好,而OSAVI的R2更是达到了0.711,RMSE达到0.64 t/hm2,RE为11.4%。依据OSAVI与地上部生物量建立的模型制作了农场水稻拔节期地上部生物量监测专题图;结论:GF-1的WFV数据在拔节期水稻地上部生物量监测中能够获得较高的精度,具有广泛的应用前景。

Objective： The estimation model to determine the aboveground biomass was established based on GF - 1 satellite remote sensing image to provide reference for water and fertilizer management and prediction for pro- ductivity at jointing stage of rice. Method： To establish and assess some high - precision estimation models, the rice - cultivation regions of Shengnong farm in Anhui Province was selected as the subject and using GF - 1 WFV satellite image to analyze correlation between biomass and spectral parameter. Result： The result showed that the rice aboveground biomass had good correlation with the spectral parameter of GF - 1 WFV and the near infrared band had a sensitive response when using single band to calculate the correlation, then the parameter was grouped by green and near infrared bands that was established by stepwise multiple linear regression, which can increase the accuracy effectively. But when using 10 common vegetable indices, the NDVI,GNDVI,RVI,SAVI EVI and OSAVI have a significant correlation with aboveground biomass, and the R^2 of OSAVI was 0.711, RMSE was 0.64 t/hm^2, RE was 11.4%. Then, the monitoring image thematic map of rice aboveground biomass was mapping by the equation which was calculated by OSAVI. Conclusion： Using GF - 1 images to estimate the biomass of rice jointing stage can provide an important basis and has broad application prospects for the precise fertilization