Automated quality inspection in assembly lines through low-cost vision system (VISION)

Important results from the project

The goal of the project was to develop a cheap and high-performance vision system for automated quality inspections based on advanced machine learning algorithms for real-time image analysis. The purpose was to enable broad use of automated quality inspections in the industry to reduce quality defects and thereby contribute to increased sustainability. The project´s objective has been fulfilled by completing a fully functioning general solution to TRL 6-7, and evaluation of this in industrial test cases shows that the purpose of the project has also been fulfilled.

Expected long term effects

The project has resulted in a fully functioning general solution on TRL 6-7. The solution is based on deep learning neural networks and can be used with cameras that cost less than SEK 300. The solution has been developed and evaluated through a number of industrial test cases and in two publicly available demonstrators. The evaluations show that the solution is very powerful and can successfully solve a wide range of different industrial problems. The assessment is that the project´s results and effects exceed what was expected at the start of the project.

Approach and implementation

The project has been conducted as a collaboration between seven partners; two universities, an institute, two large manufacturing companies, a manufacturing SME company and an industrial development center. The project work has been carried out in seven WPs where great focus has been on technical development of the solution and application in industrial test cases. Both the structure of the WPs and the cooperation in the project consortium have been very effective, and it has been particularly successful that the various parties have complemented each other in terms of competence.

External links

Offentlig webbsida för projektet