Damage Detection on Wind Turbine Blades

This research focuses on the automation of blade inspection, using different Computer Vision (CV) approaches and methods to detect damage on the wind turbine blades, making use of Convolutional Neural Networks (CNNs), and compare a model that best adapts to damage detection solution. To achieve this, different methods, architectures, and deep learning algorithms have been experimented, for Image Classification, Object Detection and Instance Segmentation tasks. An image dataset of wind turbine blade defects, annotated by experts has been used. Different performance metrics were also used to compare and evaluate each model.