Implementation of Optical Draw-In Measurement System in an Automotive Press Line: Machine Learning and Computer Vision
Translated title
Implementation of Optical Draw-In Measurement System in an Automotive Press Line
Author
Term
4. term
Education
Publication year
2024
Submitted on
2024-05-31
Pages
58
Abstract
In pursuit of product personalization, climate goals, and safety, Volvo actively invests in research on cuttingedge technologies. Leveraging validated learning and product architecture, a VT4 student from Aalborg University is dedicated to structurally developing and implementing a visionbased machine learning system. This system aims to assist in production quality control and data collection for further research. Concurrently, the student investigates whether the metal sheet draw-in during active production can be estimated using flange measurements. This report details the implementation of U-Net and YOLOv8 models, with U-Net also deployed and tested on the NVIDIA Jetson-Orin Nano edge device. The outcome is a modular product designed to aid in production quality control, although there is limited evidence supporting the capability of estimating draw-in through flange measurements.
Documents