A study was undertaken to evaluate the correlation between hyperspectral data with soil fertility status in the study area of Chinnapalem village Guntur district, Andhra Pradesh. One hudredfifty-three (153) samples were collected at different locations with GPS coordinates by adopting grid method. The hyperspectral data revealed that soil reaction, available Mg, Zn, Fe, Cu and Mn showed positive and significant correlation between VIS-NIR region of electromagnetic spectrum. However, EC and available S were negative and significant throughout VIS and SWIR. From, Stepwise regression approach the poorest fit was observed in all the properties although the highest accuracy (R2=0.467) was found for available zinc, while lowest predictability (R2=0.028) was for sand.
Read full abstract