Automatic Cell Phone Detection in Large Volume of Baggage Processing

Abstract

The research focuses on the detection of mobile phones that appear in the passenger baggage at airport arrivals. The aim of the research is to develop a method that detects mobile phones efficiently in passenger baggage at the custom scanner for the purpose to make sure that no mobile phone is passed undetected without payment of duties and taxes. It presents a machine learning based solution towards the airport security system by detecting mobile phones in a scanned image of passenger’s baggage at airport arrival. Classification is based on colour, density, size and pattern. It is challenging to ascertain if an electronic item is a cell phone or not from an x-ray image particularly when two objects are overlapping each other. The system’s performance is marred by the unavailability of high-quality x-ray images. The performance of the system increases manifolds when a high-quality image is provided as a test case. The system is able to classify the images correctly 80 percent of the time on average. The research project is of significant importance to the customs authorities as it helps them in profiling the passenger baggage at the arrival for imported mobile phones.