The Substation Intrusion Detection System is an advanced security system that aims to enhance the protection of electrical substations - critical facilities that control and monitor the electricity supplied to communities, homes, and businesses. Our system synthesizes the features of machine learning, network tunneling, protective firewalls, infrared cameras, triangulation, and data collection to create a product that can be implemented into pre-existing security systems. This innovative system strives to provide 24/7 comprehensive surveillance security and threat-detection capabilities using YOLO (You Only Look Once) for object detection, camera triangulation (Stereo Vision) for distance calculations, and virtual zones using PyTorch to establish a relationship between object boundary boxes and zones coordinates. Two OV5647 5MP camera modules are used to calculate the distance of an intruder as well as build redundancy by only displaying threats if both cameras detect a person that exceeds a confidence score of 0.5. The Substation Intrusion Detection System aims to collect and send this data securely through tunneling by establishing public and private keys via Wireguard. By utilizing these various technologies, our system’s goal is to ensure that electrical substations remain secure by minimizing the risk of damage and loss of power to communities.
Advisors: Dr. Miguel Gates and Dr. Matthew Hartmann
Following report for more detail can be found in Team 1 end of report file above