Early Identification Of Infected Tobacco Plants Using Hyperspectral Imaging Technology

The main goal of this research was to develop the real-time remote sensing system as a rapid and field based method of identifying infected plants from healthy ones at an early stage of disease development. The detection of plants disease before visibly seen by human eye can be achieved through the use of hyper-spectral imaging sensor collected data between 380-1030nm wavelengths. In this work eight pots of young Tobacco plants was used and the Black-Shank disease was inoculated to four of these pots 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 730nm and 790nm with corresponding bands of 283 and 330 was the most useful for discriminating black-shank disease severity level. The "Agricultural Stress tool" in ENVI was used and analysed the leaf moisture content from hyperspectral reflectance data. Whereby the healthy plants shows low stress sign and the weak or infected plants shows the sign of high stress.The PCA first derivatives obtained from the average reflectance show that the most important wavelength are identified at 630,680,750,790,840 and 880nm. The first three principal components from the variance curve generated they were responsible for 96% of variability of the data. The Principal Components Analysis (PCA) Scores plot shows the inoculated plants are almost entirely located in the negative area indicated inverse relationship among these two groups, whereas some scores located at the center are the medium reflectance region from both of the ports. This research indicates clearly the relationship between spectral properties and plant response in precision Agriculture.