AutoInspect

Automated Quality control of die-cast and hot-pressed metal components - AutoInspect

Purpose and goal

The project aims to develop smart, flexible and user-friendly camera systems that are capable of detecting surface defects, such as surface pores, on die-cast and hot-pressed metal components. The software of the systems developed in this project is based on machine learning algorithms. The goal is to develop solutions that, in the degree of technological maturity, are close to a possible implementation in die castings and hot pressing companies.

Expected effects and results

Expected results are that at least one type of fault on at least one article from each participating metal processing company can be detected with good accuracy under varying conditions (different brightness, angle of the article, etc.) of at least one of the camera systems developed in the project. In the short term, the effect will be that the metal processing industry realises the potential of AI-based quality inspection. In the long term, smart cameras are expected to be part of powerful, automatic production control systems that lead to a significantly reduced consumption of energy and resources and thus reduced environmental impact.

Plan and implementation

The project mainly consists of two steps. In the first step, a test bed is built at RISE, where two camera companies can collect image data and test their systems under production-like conditions. A third system based on cheap webcams and open source will also be developed and tested in this test bed. The latter system is developed by RISE, with some guidance from the University of Skövde, which in a parallel ongoing project is developing a similar system for other industries. In the second stage, tests are run in sharp production lines.