Abstract:

This study research on the feasible of using domestic HJ satellite images for rice yield forecasting in Zengcheng, where field is broken and there are a lot of cloudy or rainy days during growth stage of rice. The study conducted in Zengcheng in 2010. During field campaign, rice growth vigor during its growth stage and final yield were monitored. Meanwhile, multi-temporal remote sensing images were also collected. Rice planting area was extract, and principal components analysis methods was used to establish rice yield prediction model based on relationships among reflectance spectra, growth vigor and yield. The results showed that, the domestic HJ satellite can provide effective images, which facilitate accurately extract rice growth area and forecast gain yield. For early rice in 2010, this study got an accuracy value of 97.3% when extracting rice growth area, and has a R2 value of 0.73 and a relative error value of 12% for grain yield prediction. The results promote the application of domestic satellite in this area.