multispectral images from the orchard, detecting HLB citrus disease HLB HSI

The dataset contains a collection of 14 VisNIR bands multispectral images collected in brazillian citrus orchard. They describe Large portions of canopy approximately covering 3 m². It contains around 2800 images acquired by UNESP (https://www2.unesp.br/) over two seasons, between 2023 and 2024, in the region of Sao Paulo State, Brazil.

Images fall into two categories, HLB infected and non-infected. Two format are proposed hdf5 (.h5) and collections of TIFF images (1 per band for a total of 14 files per image).

A python based workflow to use the image is available from GitHub: https://github.com/florent-abdelghafour/MSI-DeepLearning

A description of the first part of teh dataset anbd its use in deep learning can be found in : Porto, L. R., Abdelghafour, F., Oviedo, M., Imai, N. N., Tommaselli, A. M. G., and Bendoula, R.: Use of ResNets for HLB Disease Detection on Orange Leaves Using Terrestrial Multispectral Images, ISPRS Ann. Photogramm. Remote Sens. Spatial Inf. Sci., X-3-2024, 331–337, https://doi.org/10.5194/isprs-annals-X-3-2024-331-2024, 2024. doi: DOI: 10.5194/isprs-annals-X-3-2024-331-2024