Advanced Manufacturing Technology Development Using Synthetic Data for Machine Vision Quality Inspection

The Tauber team will work with Ford and Microsoft Technical experts to develop a new CAD-based data generation and machine learning framework that uses synthetic images to train deep learning machine vision models. This enables defect detection early on in the assembly and manufacturing process immediately following production launch. 

The implementation of the models will result in time savings, reduction of repair costs and overall cost savings, quality improvement, and long-term advantages of a more flexible and generalizable computer Vision Inspection system.

The deliverables for this project include a proof of concept machine vision inspection model that focuses on electrical connectors for the parking brake and wheel speed sensor and a business case detailing the gains with successful implementation.

Student team:

Yvonne Lin – EGL (BSE & MSE in Biomedical Engineering)

Christian Zung – EGL (BSE in Mechanical Engineering & MSE in Electrical Engineering)

Project Sponsors:

Frank Maslar

Mark Goderis

Melinda Hunsaker

Faculty Advisors:

Fred Terry, College of Engineering

Lindy Greer, Ross School of Business