General Motors (GM) is a global company that has grown to become one of the largest and most recognizable automotive brands in the world. Founded in 1908 by William C. Durant, GM has established itself as a powerhouse within the industry, and has since expanded into various fields, such as Auto Loan Finances, Defense, and Software.

GM has a longstanding history of protecting natural resources, minimizing environmental impact, and preserving the biodiversity across its operations worldwide. The company is now exploring new ways to efficiently identify and monitor species at its sites.

Our Habitat Identification using Drone Imaging tool makes it easier to map and log plant species more efficiently through an intuitive web application.

When a user provides drone imagery of a habitat, the images are processed by a machine learning model that detects and classifies different plant species. The results of the model are stored in a database and displayed through an easy-to-read and interactive UI.

The application also supports continuous model learning by enabling users to confirm or correct the model’s outputs. This means the model not only improves over time but also learns to classify unknown plant species through user feedback.

Our application efficiently reads drone imagery and identifies the different plant species present within the habitat. It provides the experts at GM a tool for tracking plant species abundance and diversity over time.

Our Habitat Identification using Drone Imaging tool is built as a web application using a React together with TypeScript front end. The back end built on Flask with Python. The machine learning model is developed using Segment Anything Model in combination with the ViT to detect and segment plant species. PostgreSQL is used to store the data from the drone footage and model.