ABSTRACT
Computer theft in computer labs causes academic damage to courses
that require this resource and ends up directly harming students. In
this context, this paper describes a methodology applied to detect
computer removal through video analysis in real-time. For each
frame, image processing and computer vision techniques were used,
subtracting background information, binarization, segmentation
of the region of interest and definition of contours. The case study
was developed at a Brazilian university. For theft detection, it was
considered a black computer tower case carried by people leaving
the laboratory. Monitoring is carried out by a camera positioned in
front of the lab exit door. The software developed alerts a suspicious
activity that may indicate a possible computer theft.