Samueli School of Engineering

Abstract—SALUS is an open-source security system that uses inexpensive, affordable hardware, specifically catering towards primary and secondary schools. Students, teachers, administrators, and parents have endured the reality of school shootings. SALUS’ intent is to alleviate gun violence on school campuses and keep the institutions safe. SALUS is easily activated and managed through a graphical user interface, and the system autonomously contacts faculty, administrators, and local law enforcement if a gun is detected on camera. The system was also designed to adhere to federal FERPA laws. The user-interface is created using QT creator, and the physical working prototype is made using a Raspberry Pi 4 and a Dorhea Raspberry Pi Camera. The Raspberry Pi uses UDP(User Datagram Protocol) to send images from the camera to the desktop server which detects motion using functions from the OpenCV library. Using YOLOv3 and Google Research’s Colab, motion detected images are processed and searched for weapons. If a weapon is detected in the images, Twilio is used to send text messages and SMTP protocols are used to send emails from Google’s email service Gmail to the contacts listed on the GUI. The working prototype has achieved an accuracy of 80% or greater in non-testing conditions. The best recorded time the system took to notify those on the contacts list was 8.97 seconds. Thus, our system successfully decreases tactical time and notification time, while maximizing response time to save more lives overall.