This repository contains code for detecting and classifying host trees using Google Street View imagery. The project consists of four main components:
The /image_downloaders folder contains tools for obtaining two types of images:
- iNaturalist host tree images
- Google Street View panoramic images
In the /tree_classification folder, you'll find scripts for developing and evaluating a convolutional neural network (CNN) tailored for host tree classification.
The /tree_detection folder houses scripts to apply a pre-trained object detection model (YOLOv5) to identify trees in Google Street View panoramic images.
The /tree_geolocation folder contains scripts that integrate detection and classification models into a workflow. This workflow triangulates and assigns geolocation coordinates to street trees.
