The application of hyperspectral remote sensing for crop separability and for monitoring crop biophysical parameters : a case study of the West Nipissing Agricultural zone

Abstract

The use of non-destructive methods of gathering important agricultural information has become increasingly desirable in the last decade. Studies suggest that hyperspectral remote sensing provides a highly accurate, non-destructive and cost efficient way of monitoring crop biophysical parameters. The purpose of this dissertation is two-fold; first to determine what hyperspectral vegetation indices (VI) are best at predicting leaf area index (LAI) measured using an AccuPAR LP-80 Ceptometer, and chlorophyll content index (CCI) measured using an OPTI-SCIENCES CCM-200 chlorophyll content meter, within the 325 ? 1075 nanometer (nm) range, for five cash crops commonly grown in northeastern Ontario, Canada. Second, to determine what hyperspectral narrow wavebands are best at distinguishing between the same five local cash crops within the 400 ? 900 nm range. The study took place over a twelve week period starting on July 7, 2011 and ending on September 21, 2011. Data were collected from ten different fields that included two of each of the following crop types; soybean (Glycine max), canola (Brassica napus L.), wheat (Triticum spp.), oats (Avena sativa) and barley (Hordeum vulgare). Regression analysis (linear and exponential) was used to analyze the data collected for the LAI and CCI investigation. Stepwise discriminant analysis and bivariate correlations were used to assess the spectral separability between the crop types. The results indicate that hyperspectral reflectance data captured using a handheld device can be used to effectively predict CCI and LAI for all crops and that the top performing VIs for prediction are location and crop specific. Moreover, hyperspectral reflectance data (400 - 900 nm) can be used to effectively distinguish between the five commonly grown cash crops in this study at almost any point during the growing season. However, the optimal time for satellite image acquisition was determined to be in late July or approximately 75-79 days after planting with the optimal wavebands located in the red-edge, green and NIR regions of the spectrum. The results of this study will help improve our knowledge of precision agriculture activities in northeastern Ontario with the ultimate goal of improving crop productivity while limiting costs and impacts on the environment.