Monitoring quality of winter wheat based on the HJ satellite images

Abstract

Field surveys were conducted at regreening, jointing, heading and flowering stage in Jiangsu in 2011. The HJ satellite images were downloaded and preprocessed. Wheat grain quality was acquired at maturity. Band values of sampling sites were derived from the satellite images of four growth stages by using GPS information, and the candidate vegetation indices were calculated with the band values. The relationship between quality indices and vegetation indices was analyzed, the results show that: DVI vegetation index at heading stage had the highest correlation with grain protein content, and SAVI at heading stage had the highest correlation with sedimentation value. Both relationship reached highly significant level. The prediction model of grain protein content was built with DVI, and the model was y = −0.003x+19.67, and the determination coefficient and RMSE were 0.376 and 1.434. Sedimentation value prediction model was built with SAVI, and the model was y=−47.13x+72.61, the determination coefficient and RMSE were 0.367 and 5.101. Thematic maps of grain protein content and sedimentation were made based on the satellite images at the heading stage and prediction model above.