The article presents a vision-based approach for automatic airport pavement inspection using Unmanned Aerial Vehicles (UAVs) and deep learning. The researchers used an optimised implementation of EfficientNet feature extraction and Feature Pyramid Network segmentation to identify pavement distress in images captured by UAVs. To overcome the lack of annotated data for training, they developed a synthetic dataset generation methodology to extend available distress datasets. The study found that using a mixed dataset of synthetic and real training images yielded better results in real application scenarios.
Publication date: 12 Jan 2024
Project Page: unknown
Paper: https://arxiv.org/pdf/2401.06019
