An AR-Assisted Deep Learning-Based Approach for Automatic Inspection of Aviation Connectors

The mismatched pins inspection of the complex aviation connector is a critical process to ensure correct wiring harness assembly, which existing manual operation error-prone and time-consuming. Aiming fill this gap, article proposes an augmented reality (AR)-assisted deep learning-based approach tackle three major challenges in inspection, including small detection, multipins sequencing, visualization. First, proposed spatial-attention pyramid network extracts image features multilayers searches for their spatial relationships among images. Second, based on cluster-generation sequencing algorithm, these detected are clustered into annuluses expected layers numbered according polar angles. Finally, AR glass as visualization platform, highlights interface warn operators automatically. Compared with other methodologies, experimental result shows that can achieve better performance accuracy support operator's efficiently.