Detecting foreign objects online.

There is a continuing demand for improving techniques that can detect foreign objects online in foods. X-rays have been used for this purpose for years, but they have a number of limitations. Research in Europe is leading to the development of a system capable of capturing and processing two orthogonal (at right angles) X-ray images of trays and individual jars of product in order to detect glass and metal.

Researchers are developing a number of technical modules that will be integrated to form an experimental rig capable of testing in a food processing facility. The modules include an X-ray sensor, image processing software, image acquisition hardware and a prototype test rig.

An X-ray sensor is being developed that will improve speed and performance, giving an enhanced signal-to-noise ratio. (The level of noise indicates the amount of background interference.) A three-dimensional (3D) software model allows you to visualize a defect and a jar from two different angles. This technique also covers two jars (side by side) so that the relationship between the different overlapping edges of the containers can be readily seen. The design of the image processing hardware is underway, as is the fabrication of a 3D test rig.

Research has shown the limitations of conventional 2D X-ray systems for detecting small pieces of glass in individual jars. It's also shown the total inability of such systems to detect glass inside several jars, such as on a tray. This packaging format is likely to increase in importance. With two views, you can overcome the orientation problems associated with a defect (e.g., glass) lying with a thin cross-section parallel to the X-ray beam.

Further information. M. Graves, Transfer Technology Ltd., 26 Roman Way, Coleshill, Birmingham B46 1HQ, England, U.K.; phone: [phone]; fax: [phone].