The increasing adoption of Artificial Intelligence (AI) is reshaping quality control in critical sectors. Within the AVIS project, this study presents a benchmark of AI-based automated visual inspection systems. Ten commercial solutions were analysed through a systematic requirement assessment, and the system was validated using a dedicated dataset. The aim was to identify the solution that best meets requirements while supporting operators during inspection. A case study with Avio Aero enabled the design and deployment of the system in a real production environment. Focused on detecting defects in aeronautical power transmission components, the system ensures high quality standards and contributes to Zero Defect certification. Its On-Premises/Edge architecture supports local processing, reducing latency and improving security. The implementation followed five phases, from data collection and model training to real-time testing. Results confirm the effectiveness of the system in defect recognition and validate the methodology as decision-making tool for industrial technology adoption.
Corallo et al. (Thu,) studied this question.