Hyperspectral imaging system sorts nonferrous materials

A team of Irish and Spanish researchers has developed a hyperspectral image processing system that could be used to sort nonferrous metals.

The researchers, from Tecnalia Research and Innovation (Zamudio, Spain) and Dublin City University (Dublin, Ireland), developed the system after determining that a system based on color identification would not be sufficient to determine the characteristics of the nonferrous metals.

The system itself consists of several modular components that include hyperspectral image acquisition and nonferrous material classification systems, and mechanical subcomponents such as a vibratory delivery system, a conveyor, and a pneumatic sorting mechanism.

The experimental results reveal that the hyperspectral machine vision system would be able to process nonferrous shredded scrap with a classification accuracy of 96.87 per cent.

The researchers detailed their research in the Jan–Mar 2012 issue of the Journal of Electronic Imaging. A complete research paper describing the design of the system can be found here.

-- by Dave Wilson, Senior Editor, Vision Systems Design