Enhancing Inspection Tasks: A Dataset for Corrosion Defects in Pipelines

Abstract

Inspection plays a crucial role in ensuring the longevity, security, and dependability of critical public infrastructure for both governments and businesses. However, traditional inspection processes are often labor-intensive and pose various risks. Consequently, there is a growing need for automation in such tasks. This research paper presents a comprehensive dataset that can be utilized to develop algorithms and systems for automating the inspection process, a critical area in the field of computer vision. The dataset encompasses a diverse range of inspection scenarios and serves as a valuable resource for advancing automation technology specifically for the inspection of steel pipes to detect corrosion defects. Real-life pipe maps have been used to derive scenarios that represent varying levels of corrosion. By leveraging this dataset, researchers and practitioners can contribute to the development of more efficient and accurate automated inspection systems, thus greatly improving the overall efficiency and long-term safety of infrastructure inspection.