Classifying Tree Genera Based on Bark Texture v1 Computer Vision Project

The purpose of this model is to identify several tree genera commonly found throughout the region of North America based on the physical texture of their bark. However, this model does not intend to identify the individual species of the trees within its dataset. Rather, its focus is on classifying the trees into four common genera: Quercus (“oak”), Fagus (“beech”), Pinus (“pine”), and Fraxinus (“ash”). The dataset consists of 200 images of bark from various tree species, formatted to include only a section of the trunk’s bark with background features cropped from the image. In further renditions of the model, more images will be added to the dataset. The images used for the dataset were accumulated from two online sources: Kaggle.com, from a project titled Photos d'écorces d'arbres du Québec - BarkNet and iNaturalist.org, from various user observations uploaded to the website.

This model serves to aid a wide range of individuals and professionals. Ecologists or researchers interested in studying the diversity of specific tree species may find this model useful for identifying tree genera while surveying a region. Arborists, foresters, and tree climbers–such as those who perform limb trimming–can also utilize this model as a tool for identifying tree genera of interest and thus educating the way they should move forward in maintenance, research, or identifying nonnative species. In addition, this model serves as a valuable public resource for various recreational and citizen-science purposes.

Note: This model was created for a class assignment in AI and Natural History at St. Mary’s College of Maryland.