Translator Disclaimer
10 October 2018 Temporal and spatial aggregation of the normalized difference vegetation index for the prediction of rice yields
Author Affiliations +
Abstract
In recent years, the Normalized Difference Vegetation Index (NDVI) has been used to help in the analysis of plant productivity, especially for rice crops. In this research, we analyze time series of NDVI (2007 to 2015) for Bangladesh to predict crop yields. A key ingredient is the rice classification of the fields. The crop yield estimations are made using rice masks and pixel-based season alignment. Furthermore, the pixel-based growing seasons are aggregated to district level, to correlate with national yield data. NDVI ~ Yield models were trained with data from 2007-2013. District specific regression models provide model fits of Adjusted R2 = 0.6 ± 0.3, estimating ricle yield with a Root Mean Square Error (RMSE) of 0.09 ± 0.05 tons/ha. Model validation with data from the results between 2014 and 2015 in rice yields estimates with prediction errors of 14.7%. In conclusion, we show with this research that the method of aggregation of NDVI temporally as well as spatially can lead to improving correlation and can predict rice yields.
© (2018) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
W. Suijker and E. Aparicio Medrano "Temporal and spatial aggregation of the normalized difference vegetation index for the prediction of rice yields", Proc. SPIE 10783, Remote Sensing for Agriculture, Ecosystems, and Hydrology XX, 107830Z (10 October 2018); https://doi.org/10.1117/12.2319189
PROCEEDINGS
12 PAGES


SHARE
Advertisement
Advertisement
Back to Top