A Human Intruder Detection System for Restricted Sensitive Areas

The development of an automated intruder detection system for restricted sensitive areas or unattended border areas is a major concern for national security perspective to avoid illegal entries of outsiders to those areas. Due to the larger radius of those areas, it is quite difficult to keep track of the activities through a manual human tracking process at regular intervals for periodic monitoring. Also, the safety of the people involved in the manual supervision process by visiting the sites is of major concern while no prior information is available on any intruder already entered that place. This work focuses on the development of an automatic human intruder detection system that can detect illegal entries of humans in restricted areas and can report to the control room in time. The model presented uses the YOLO Model to detect human objects in videos captured by cameras with some standard resolution. A number of experiments were conducted to evaluate the performance of the model under different contexts. The results obtained show the efficiency of the model in detecting human objects over other objects in the captured videos with an accuracy of approximately 96% in the considered environmental conditions.