Identification of optimal hyperspectral wavelength to discriminate maize and sorghum crops using AVIRIS-NG data in Guntur district, Andhra Pradesh, India
SK Tiwari and V Raghu
In this study, we utilized AVIRIS-NG hyperspectral data captured during the Rabi season in February 2018 in a region of Guntur District, Andhra Pradesh, where Maize and Sorghum crops are extensively cultivated. Despite having parallel crop calendars and spectral similarities, distinguishing between Maize and Sorghum crops during similar vegetative growth stages poses a challenge. The high spectral resolution of AVIRIS-NG data allows for precise identification of subtle changes in the objects under study.
The key finding of our research is the identification of sensitive spectral bands within specific wavelength regions that facilitate the differentiation of these two visually similar crops. Our results indicate that the spectral range from 1553 nm to 1749 nm of AVIRIS-NG data provides optimal discrimination between Sorghum and Maize crops, irrespective of their growth stages. Notably, the wavelengths at 1649 nm and 1654 nm emerge as particularly suitable for distinguishing between Maize and Sorghum crops, as indicated by statistical separability measures such as Wilks' Lambda and F-Value. At these wavelengths, we observed significant results with Wilks' Lambda values of 0.388 and 0.387, and F-Values of 39.29 and 39.46, respectively, further supporting their efficacy in crop discrimination.
SK Tiwari, V Raghu. Identification of optimal hyperspectral wavelength to discriminate maize and sorghum crops using AVIRIS-NG data in Guntur district, Andhra Pradesh, India. Int J Res Agron 2024;7(3S):187-194. DOI: 10.33545/2618060X.2024.v7.i3Sc.429
