INVESTIGATION OF A CLASSIFICATION-BASED TECHNIQUE TO DETECT ILLICIT OBJECTS FOR AVIATION SECURITY

摘要

In this paper we present an initial investigation into the use of a classification-based technique for illicit object detection in aviation security. Current threats in aviation security are becoming more sophisticated in that it is extremely difficult to detect possible threats of terrorism without severely hindering passenger life style. In order to provide adequate security, previous work by the authors has proposed an intelligent security technology framework to provide the civil aviation authority with maximum security whilst minimising adverse impacts on airlines and airport operations. In this work, the feasibility of employing a classification-based technique is investigated for the purpose of identifying illicit material in hand luggage. In this research, a neural network trained with back-propagation is used in conjunction with a newly proposed feature extraction technique for classifying various object images. Encouraging results are reported that may facilitate future, automated hand luggage scanning.