GoalKeeper ESBBD : A European Spruce Bark Beetle Detector

Sammanfattning

This thesis investigates the development of a new detection method directed towards Spruce Bark Beetle infestations in Spruce trees, and the evaluation of this method’s achieved performance in terms of accuracy and cost effectiveness. Spruce Bark Beetles have the capacity to devastate vast regions of woodland. Continuous early detection is vital for removing small scale infestations, before devastating mass scale outbreaks occur. The established methods for detection have several drawbacks, ranging from delayed detection times to not capturing all symptoms. The proposed detection system to counter this, consists of a drone capturing below-canopy footage, which is then streamed to a server performing object detection, that transmits the processed frames to a connected client device. The proposed benets over the established detection methods are: decreased prediction time allowing for an increased inspection coverage, increased diagnostic accuracy, and increased automation. Despite several technical and practical challenges, the project resulted in a functional proof of concept with an achieved detection accuracy of 63%. The results indicate that integrating the GoalKeeper ESBBD system with existing inspection methods can improve the speed, coverage, and reliability of early Spruce Bark Beetle detection in forest environments.