Title:	AI for Wildflower Identification and Classification
DNr:	Berzelius-2024-465
Project Type:	LiU Berzelius
Principal Investigator:	Derya Akbaba <derya.akbaba@liu.se>
Affiliation:	Linköpings universitet
Duration:	2024-11-27 – 2025-06-01
Classification:	10204
Homepage:	https://github.com/sarakarimi/PlantMap
Keywords:

Abstract

Identification of wildflowers is critical for understanding the health and composition of a local ecosystem. One of the main challenges with creating an accurate roster of wildflowers is that they grow at different rates, grow across a large area, require expertise to identify, and often small in nature. There is very limited prior research in this area and the work that does exist, does not use state of the art ML/AI models. So with this work we propose to use the Eindhoven Wildflower Dataset (EWD) to train state of the art image models (ex. CLIP) to classify drone images of wildflowers from the Norrland region of Sweden. The work will first be submitted for a WASP class project and then submitted as a publication.