Abstract

The main goal of this research was to develop the real-time remote sensing system as a rapid and field based method of identifying healthy and infected plants at an early stage of disease development, before visibly seen by human eye. This can be achieved through the use of hyper-spectral imaging collected data between 380 to 1030 nm wavelengths. The black-shank disease was inoculated to tobacco plants as a model system for testing this technology. The hypercubes images acquired was processed using ENVI software and the “unscrambler” statistical analysis software for principal components analysis (PCA). Spectral parameter of reflectance sensitivity was used to find the optimal wavelengths for determining and evaluating the level of damage by the black-shank fungus. The result of this research shows that, the spectral reflectance decreases significantly with the increasing severity level in both the visible and near-infrared wavelength ranges. Also the wavelength of 730 and 790 nm with corresponding bands of 283 and 330 was the most useful for discriminating black-shank disease severity level. This research indicates clearly the relationship between spectral properties and plant response.

Key words: Hyperspectral, image processing, moisture, reflectance, tobacco plants, black-shank.