Metadata for Automated plant species identification from the stomata images using deep neural network

Published: 31 July 2023| Version 4 | DOI: 10.17632/4brcwhmvyk.4


Our study has been conducted at the Sundarbans Mangrove Forest and Ratargul Fresh Water Swamp Forest. Both study sites have distinct ecological characteristics. While the Shundarbans Mangrove has a complex network of tidal waterways and is regularly inundated by saline water, the Ratargul is linked with the Gowain River and faces seasonal flooding with the fresh water from the Gowain River during the rainy season, which is generally from May to early October. There is a stomata dataset of 11 species collected from these two study sites: 9 species collected from the Sundarban mangrove forest and 2 species collected from the Ratargul Swamp Forest in Bangladesh. These are the species worth mentioning in the two forests. Hence, we used this dataset for automated species identification. Sundarban mangrove species- 1. Aglaia Cucullata 2. Bruguiera gymnorrhiza 3. Bruguiera sexangula 4. Cerbera manghas 5. Ceriops decandra 6. Excoecaria agallocha 7. Heritiera fomes 8. Sonneratia apetala 9. Xylocarpus moluccensis Ratargul swampforest species- 10. Barringtoni aacutangula 11. Pongamia pinnata For more please read metadata file "Stomata ISO19139 Metadata.xml"


Steps to reproduce

The leaves collected from two forest zones were preserved in plastic zipper bags to maintain the leaf moisture and the heavy layer of accumulated salt on the leaf surface due to saline water transpiration helped avoid any fungus attack. Later, stomata images were taken from leaves using the nail polish imprint method. Later, the imprints were analyzed under an electronic light microscope with 8 MP scope-mounted camera (MU800) and photos were taken using both 10x and 40x lenses.


Shahjalal University of Science and Technology Department of Forestry and Environmental Science


Plant Physiology, Freshwater, Mangrove Swamps, Stomatal Conductance