Year
1990
Abstract
An important aspect for the overall security in a nuclear f a c i l i t y is to detect the introduction of contraband that can be used to damage the f a c i l i t y or the people in it. Contraband detection is commonly done by metal detectors and X-ray scanning machines. Although there have been several technological advances in the area of explosives detection, the problem of weapon detection is still very dependent on human vigilance. This paper will present a study of an automated weapon detection system based on X-ray image processing. Given reference images of certain types of contraband, the system first extracts their unique features and stores them in a database. Once the database is established, images from an X-ray scanning machine are checked to determine whether any clutter in the image can possibly contain a given contraband. The system then alerts the operator with a video display of those threat clutters. The implementation and operation of the system will be discussed in detail.