| Crop residues left on agricultural field after harvest is an effective alternative, among others, to minimize the harmful effects of wind and water erosion, to increase the quantity of nutriments in the soils and to reduce CO2 emissions in the atmosphere. During this master's degree, two methods used in remote sensing were applied to the Probe-1 hyperspectral image of an agricultural field situated in the Southeast Saskatchewan, crop residue index and neural networks. The purpose was to determine which of the two methods was the most effective to accurately map and estimate crop residues. It is worth while mentioning that this the first time hyperspectral data were used to mapping ends of crop residue, which constitutes an advancement in the domain. To complete the dataset, several spectral reflectance measures were taken from different types of crop residues (corn, wheat, herb, soya and sunflower) and different types of bare soils, obtained by the spectroradiometer GER3700. Nine crop residue indices were used (BI, CAI, NDI-1, NDI-2, SACRI-1, SACRI-2, MSACRI-1, MSACRI-2 and CRIM). (Abstract shortened by UMI.)... |
