Machine learning-based wood anatomy identification: towards anatomical feature recognition

What is it about?

Wood identification is important for combating illegal logging and enforcing timber regulations. Traditionally, trained wood anatomists examine wood samples under a microscope to identify key anatomical features and determine the wood's species. However, there are too few experts to handle the volume of wood testing required. New technologies can help._x000D_ _x000D_ This study reviews how Artificial Intelligence (AI) methods like Machine Learning (ML) are being applied to wood identification. Machine learning algorithms are trained on thousands of wood sample images to recognize the anatomical patterns of different tree species. These ML models can then analyze new wood samples and predict their species._x000D_ _x000D_ Some methods try to replicate how anatomist identify wood by detecting the same anatomical traits humans look for, like vessel arrangement or ray width. Other methods take a computer vision approach, training algorithms to recognize wood species based just on the overall appearance and patterns in the images. Each methodology has strengths and weaknesses. _x000D_ _x000D_ This paper reviews the current state-of-the-art in this emerging ML wood identification field and identifies key opportunities and future challenges. It will be a helpful resource for researchers aiming to develop innovative ML solutions for wood identification and combating illegal timber trade._x000D_

Featured Image