Dataset for tomato bacterial wilt severity identifiction
Description
This dataset is the canopy hyperspectral data of tomato plants at different bacterial wilt levels. The hyperspectral data were acquired by using a Specim IQ camera (Specim Ltd., Oulu, Finland). Tomato bacterial wilt severity is usually assessed by visual inspection of the percentage of wilted leaves. In this experiment, it is classified as: level 0, healthy; level 1, infected but no obvious symptoms of wilting; level 2, infected and 1-33% of leaves wilting; level 3, infected and 34-66% of leaves wilting; level 4, infected and 67-100% of leaves wilting. The dataset has 293 samples, of which 120 are at level 0, 68 are at level 1, 39 are at level 2, 33 are at level 3, and 33 are at level 4. Each sample contains two data, one is the average spectrum of the region of interest (ROI) in the canopy hyperspectral image, and the other is the 500 pixel-level spectra randomly selected from the ROI in the canopy hyperspectral image. The data is given in .mat format, which can be processed directly in Matlab and Python.