Hyperspectral remote sensing for monitoring growth of rice (Oryza sativa) crop

A field experiment was conducted with 3 cultivars of rice (Oryza sativa L.), viz. ‘PR 114’, ‘PR 116’ and ‘PR 118’, under 5 levels of nitrogen (0, 75, 125, 175 and 225 kg N/ha) during the rainy (kharif) seasons of 2010 and 2011 at Ludhiana, Punjab. Growth parameters like leaf-area index (LAI), chlorophyll content, number of tillers, chlorophyll-concentration index (CCI) and plant height were recorded. Spectral indices such as normalized difference vegetation index (NDVI), ratio-vegetation index (RVI), moisture-stress index (MSI), green index (GI), leafchlorophyll index (LCI) and plant-senescence reflectance index (PSRI), were computed using the multiband spectral data, recorded with Fieldspec®Pro 2000. The spectral indices at booting stage had the best correlations with the crop parameters and had ‘R²’ values between 0.68 and 0.85. Spectral indices correlated well with grain yield (R² = 0.59–0.85) and 1, 000-grain weight (0.64–0.84). But the best correlations were recorded with NDVI and R² = 0.82 for grain yield, R² = 0.81 for 1, 000-grain weight, and with LCI, R² = 0.85 for grain yield, R² = 0.84 for 1, 000grain weight. The leaf colour index (LCI) and NDVI were found to be the best indices among the selected and can be used to predict plant-growth parameters and yield at the booting stage of rice crop.