Project-AgML/tropical_forest_classification

Tropical Forest Classification

This dataset provides real RGB images of tropical forest tree species, collected in a laboratory environment in Popayan, Colombia. Images were captured using a handheld digital magnifying glass at 640x480 pixel resolution, focusing on detailed botanical specimens for classification tasks. The dataset contains raw and augmented versions.
The raw dataset contains 10,795 images.
Images per class:

• Campnosperma panamensis: 826
• Cedrela Odorata: 1,128
• Cedrelinga cateniformis: 1,189
• Cordia alliodora: 929
• Cupresus Lusitanica: 815
• Dialyanthera gracilipes: 1,100
• Eucalyptus globulus: 1,105
• Fraxinus uhdei: 1,025
• Handroanthus chrysanthus: 1,106
• Humiriastrum procerum: 1,001
• Pinus patula: 571

The augmented dataset contains 9,664 images.
Images per class:

• Campnosperma panamensis: 823
• Cedrela Odorata: 0
• Cedrelinga cateniformis: 1,189
• Cordia alliodora: 929
• Cupresus Lusitanica: 815
• Dialyanthera gracilipes: 1,100
• Eucalyptus globulus: 1,105
• Fraxinus uhdei: 1,025
• Handroanthus chrysanthus: 1,106
• Humiriastrum procerum: 1,001
• Pinus patula: 571

This dataset is indexed on https://project-agml.github.io/ as part of the AgML python library.

Citation

<!-- TODO: add BibTeX citation -->

The dataset itself can be cited as:

Ordoñez Urbano, C. F. (2020). Tropical Forest Species [Dataset]. Mendeley. https://doi.org/10.17632/YZZCBYVGMH.3

This dataset was reformatted from its original format to match HuggingFace standards.