Mangrove Tree Propagules Classification via Deep Learning

Examining a machine learning model that can predict the type of mangrove species, from the images of the different mangrove tree propagules, is the objective of this study. Dataset used here were downloaded online. For propagule classification tool, transfer learning using Deep Network Designer app was used. Convolution Neural Network (CNN) based on the inception architecture was used where GoogleNet model was the exact design. Global warming is already happening, and these trees play important roles in lessening the impact of climate change. Classifying what type of mangrove trees to plant according to proper mangrove zonation will help a lot and will give a greater fighting chance to protect coastlines and people during extreme weather conditions and a whole lot more because these trees can also give food, medicine, timber, and habitat for marine species, and other animals. During the classification experiment/validation, an average of 99.7% accuracy was achieved.