The major objectives of this study were to identify spectral characteristics associated with rice yield and to establish their quantitative relationships. Field experiments were conducted at Shi-Ko experimental farm of TARI’s Chiayi Station, during 2001 to 2005. Rice cultivar Tainung 67 (Oryza sativa L.), the major cultivar grown in Taiwan, was used in the study. Various levels of rice yield were obtained via nitrogen application treatments. Canopy reflectance spectra were measured during entire growth period and dynamic changes of characteristic spectrum were analyzed. Relationships among rice yields and characteristic spectrum were studied to establish yield estimation models suitable for remote sensing purposes. Spectrum analysis indicated that the changes of canopy reflectance spectrum were least during booting stages. Therefore, the canopy reflectance spectra during this period were selected for model development. Two multiple regression models, constituting of band ratios (NIR/RED and NIR/GRN) were then constructed to estimate rice yields for first and second crops separately. Results of the validation experiments indicated that the derived regression equations successfully predicted rice yield using canopy reflectance measured at booting stage unless other severe stresses occurred afterward.
We also integrated multiple regression models, derived from reflectance spectrum measurements and using band ratios (NIR/RED and NIR/GRN) as independent variables, with SPOT 5 multispectral images taken at booting stage to predict rice yield before harvest. A 4.8-ha paddy rice field was used as testing ground for the accuracy of prediction with the rice yield prediction model. Within the site, different rice yield scenarios were produced by using combinations of rice varieties, Japonica and Indica type, nitrogen rate and drought treatments. Rice yields harvested in 10m X 10m mesh were used as ground truth data for comparison. The regional rice yield map is produced with the rice yield prediction model using SPOT 5 images taken at booting stage in this study. The results from the regional rice yield map shows that the relative errors between actual yield and predicted yield in the first season and second season in 2014 are lower than 5%. Those have demonstrated its potential for using SPOT 5 images to estimate the regional rice yield with the rice yield prediction model derived from reflectance spectrum measurements and using band ratios (NIR/RED and NIR/GRN) as independent variables.