Automating Remanufacturing Inspection using Robotics & Computer Vision

Institut

Institut für Werkzeugmaschinen und Betriebswissenschaften (TUM-ED)

Typ

Bachelorarbeit / Semesterarbeit / Masterarbeit /

Inhalt

experimentell /  

Beschreibung

Background:

In remanufacturing, used products are restored to like-new condition. To do so, they are first disassembled into their components, which are then cleaned, and inspected to determine whether they can be reused or need to be reworked or discarded. This assessment is therefore an important process step. Today, this is still done manually in most cases, due to the many different types of defects and product variants. Modern deep learning methods combined with robot-based image acquisition show high potential to automate such inspection processes.

Objective:

The objective of the thesis is to investigate approaches for automated part inspection in remanufacturing. This includes planning suitable poses and trajectories for robot-based image acquisition as well as developing deep learning solutions for automated defect detection (i.e., classification and/or segmentation methods). Ideally, the developed methods should be able to handle a high number of varying defect patterns and product variants.

Skills:

• Interest in robotics, computer vision, and deep learning topics
• Programming experience, e.g., in Python
• Experience with ROS or PyTorch is a plus

 

Please apply via email, incl. a CV and transcript of records, to johannes.bauer@iwb.tum.de

Möglicher Beginn

sofort

Kontakt

Johannes Bauer
Raum: 1332@5503
johannes.baueriwb.tum.de

Ausschreibung