A Review of Super-resolution in Industrial Applications: Methods, Task Challenges, and Future Trends

Super-resolution, a pivotal image enhancement technique, improves the resolution of low-quality images, enabling greater visual detail and clarity. Despite extensive research on single-image super-resolution, existing reviews predominantly emphasize theoretical advancements, with limited attention to the practical challenges of deploying super-resolution methods in industrial fields. Furthermore, mainstream super-resolution datasets often diverge from the complexities of industrial environments, resulting in suboptimal model adaptation and performance. This paper presents a literature review of super-resolution methodologies tailored for industrial applications, focusing on real-world image enhancement and their integration with downstream tasks. Key advancements in super-resolution network architectures and their applicability to downstream tasks like quality inspection, detection, and process monitoring are thoroughly examined. A descriptive comparative overview of state-of-the-art methods offers a preliminary reference for researchers and practitioners. Finally, future research directions are outlined to address existing gaps and explore the future work on the development of robust super-resolution solutions for industrial applications.