Title :
Predicting grain starch content of winter wheat through remote sensing method based on HJ-1A/1B images
Author :
Tan, Changwei ; Junchan Wang ; Guo, Wenshan ; Wang, Jihua ; Huang, Wenjiang
Author_Institution :
Jiangsu Province Key Lab. of Crop Genetics & Physiol., Yangzhou Univ., Yangzhou, China
Abstract :
The purpose of this study is to further improve the accuracy of predicting winter wheat grain quality with remote sensing, and to enhance the prediction mechanism. In order to predict grain starch content (GSC) in winter wheat using HJ-1A/1B images, The experiment was carried out in Jiangsu regions during 2010 winter wheat growth season. Based on HJ-1A/1B image, synchronous or quasi-simultaneous ground observations of leaf nitrogen content (LNC) and grain quality parameters of winter wheat under different periods. Firstly, this study analyzed the relationships between GSC and LNC, and between LNC and satellite remote sensing variables. Secondly, based on remote sensing variable and LNC, the quantitative relationship models were established to predict GSC, and then evaluated with independent samples. Finally, the indirect model of predicting GSC based on remote sensing variable and LNC was compared to the direct model based on only structure insensitive pigment index (SIPI). The results showed that: at the booting stage, winter wheat GSC had a significantly positively correlation with Band (B1), Band4 (B4), SIPI and plant senescence reflectance index (PSRI), and then it also had a significantly association with nitrogenous reflection index (NRI). At last, a direct model for predicting GSC was established with only SIPI. At the same time, LNC in this period also showed a higher correlation with GSC. Based on the high relationship between LNC and SIPI, an indirect model of predicting GSC also was established. The indirect and direct models were evaluated with independent samples by the determination coefficient (R2) with 0.662 and 0.533, the root mean square error (RMSE) with 5.72% and 6.34%, respectively. The indirect model based on SIPI and LNC performed better to predict wheat GSC than the direct model based on only SIPI, and obtained the higher accuracy by 9.7% than the direct model. It is concluded that the research can provide an effective way to impr- ve the accuracy of predicting wheat quality based on aerospace remote sensing, and contribute to large-scale application and promotion of the research results.
Keywords :
agricultural products; artificial satellites; geophysical image processing; mean square error methods; pigments; proteins; reflectivity; remote sensing; GSC; HJ-1A/1B images; Jiangsu region; LNC; aerospace remote sensing; determination coefficient; grain starch content prediction; leaf nitrogen content; nitrogenous reflection index; plant senescence reflectance index; quasisimultaneous ground observation; root mean square error; satellite remote sensing variables; structure insensitive pigment index; synchronous ground observation; winter wheat grain quality; Agriculture; Earth; Nitrogen; Predictive models; Proteins; Remote sensing; Satellites; Grain starch content; HJ-1A/1B image; Leaf nitrogen content; Prediction model; Winter wheat;
Conference_Titel :
Artificial Intelligence, Management Science and Electronic Commerce (AIMSEC), 2011 2nd International Conference on
Conference_Location :
Dengleng
Print_ISBN :
978-1-4577-0535-9
DOI :
10.1109/AIMSEC.2011.6009646