Usefulness of in-situ hyperspectral data to develop prediction model of dual purpose Sorghum (Sorghum bicolor (L.) Moench) grown under semi-arid condition of India
2014
Deb, D. | Singh, J. P. | Rai, S. K.
Crop yield forecasts are extremely useful in formulation of policies regarding stock, distribution and supply of agricultural produce. In this paper, an effort was made to develop regression models for biomass and seed yield production of dual purpose sorghum (Sorghum bicolor) incorporating canopy spectral reflectance indices collected during three different growth stages of the crop in 2011 and 2012. The results showed that the correlation between the reflectance at each wavelength and the aboveground biomass has the maximum negative correlation coefficient (r= - 0.21, p< 0.05) at 690nm. The r values changed sharply from 690nm to 750nm. The maximum positive correlation coefficient (r=0.61, p<0.05) comes around 800nm in the Near Infra-Red (NIR) portion. In order to determine the plant stages more appropriate for yield forecasting, the indices more sensitive to yield variations, pearson correlation coefficient was calculated between the indices and grain, fodder yield for each sampling date. It was found that at flowering stage, almost all the indices were highly sensitive to both grain and fodder yield variations. Different narrowband vegetative indices like Normalized Difference Vegetation Index (NDVI) -based VIs (NDVI₁, NDVI₂, NDVI₃ and NDVI₄) and Ratio-based VIs (RVI₁ and RVI₂) were calculated to estimate the above ground biomass. Applying linear and non-linear regression approach it was found out that for both dry fodder yield and grain yield could be well estimated using non-linear function using the ratio indices i.e., X=RVI₂(735,706) and the root mean square error (RMSE) is minimum and is maximum for this function.
Show more [+] Less [-]AGROVOC Keywords
Bibliographic information
This bibliographic record has been provided by National Agricultural Library