Robust Wind Turbine Blade Segmentation from RGB Images in the Wild

TL;DR

This paper introduces a robust, data-driven segmentation algorithm for wind turbine blades in RGB images, enhancing U-Net with a specialized loss and additional steps for improved accuracy and reliability in real-world conditions.

Contribution

The authors develop a novel segmentation method combining a tailored loss function with image preprocessing and post-processing steps, improving accuracy and robustness over existing approaches.

Findings

Achieved 97.39% accuracy in blade segmentation

Enhanced U-Net performance with a custom loss and regularization

Implemented a multi-step process for reliable segmentation in wild images

Abstract

With the relentless growth of the wind industry, there is an imperious need to design automatic data-driven solutions for wind turbine maintenance. As structural health monitoring mainly relies on visual inspections, the first stage in any automatic solution is to identify the blade region on the image. Thus, we propose a novel segmentation algorithm that strengthens the U-Net results by a tailored loss, which pools the focal loss with a contiguity regularization term. To attain top performing results, a set of additional steps are proposed to ensure a reliable, generic, robust and efficient algorithm. First, we leverage our prior knowledge on the images by filling the holes enclosed by temporarily-classified blade pixels and by the image boundaries. Subsequently, the mislead classified pixels are successfully amended by training an on-the-fly random forest. Our algorithm demonstrates its effectiveness reaching a non-trivial 97.39% of accuracy.