Artificial Intelligence Applied to Trees "Fingerprints" Identification in the Amazon Rainforest.

In the Amazon rainforest, trees are often given the same name because they share similar external characteristics such as color, smell, texture, and durability. Scientists face difficulty in identifying the unique features of these trees using a magnifying glass due to the similarity in their traits. To overcome this challenge, we propose using image classification models embedded in smartphones that can identify trees without requiring an internet connection. These models were trained on a dataset of 30 tree species using images that were obtained through macroscopic analysis of the tree’s structure, enabling precise identification. The accuracy rate of this approach was over 95%, demonstrating its effectiveness.