As the global threat of terrorism continues to escalate, finding efficient ways to ensure the safety of the public is becoming a major concern for the authorities. This paper presents an investigation of scanning and detection of concealed weapons with possible applications in high risk areas like airports. Using a passive and non-intrusive scanning method like Infrared (IR) imaging, and combining it with a visual image, we hope to devise a scheme which will be able to highlight sufficiently the presence of a concealed weapon but also protect the privacy of the person that is being scanned. As an initial phase of the investigation, this work is mainly based on various image processing and computer vision techniques. They are including: image registration, image fusion and image segmentation. The experimental results are shown to be promising but there are some limitations in terms of hardware used, adopted techniques and implementations. Some improvements including using more advanced imaging sensor, employing feature based image registration and segmenting images using advanced clustering methods could be addressed. The work is useful for human operators to determine weapons and non-weapons with the minimum false detections.
展开▼
