Robust Wind Turbine Blade Segmentation from RGB Images in the Wild
This work addresses a domain-specific need for robust visual inspection in the wind industry, offering an incremental improvement over existing methods.
The paper tackles the problem of automatically segmenting wind turbine blades from RGB images for maintenance, achieving 97.39% accuracy with a novel algorithm that enhances U-Net using a tailored loss and additional processing steps.
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.