CSIS Ecosystem: Impacts of Pragmatic Detection on a Campus Surveillance Case Study

Research Context: In modern universities, where there is a constant flow of people, it is essential to implement computer vision surveillance systems to detect incidents and alert security personnel, in order to ensure the safety of all members of the academic community. Practical Problem: Universities are faced with numerous security issues, including limited resources, a diverse range of technical infrastructure, and a mix of old and new systems. Besides, traditional security methods are reactive, slow to notice incidents, which in turn makes quick response in emergencies very hard. Proposed Solution: We present the solution in the form of the Campus Surveillance Interoperability System (CSIS), which is an open-source, modular, and interoperable architecture that we put forth for the specific surveillance needs of universities. CSIS utilizes a variety of computer vision models from the YOLOv11x set, specialized in recognizing weapons, fires, floods, graffiti, suspicious behavior, and license plates. Related IS Theory: Our paper adopts a Socio-technical theory, which recognizes that the effectiveness of information systems depends on the interaction between social and technical components rather than on technology alone. Research Method: In a real-world university setting, we put our proof of concept to the test. We had real-time video stream input, ensemble-based inference, and a centralized alert management system. We assessed the accuracy of the models. Summary of Results: We achieved an average accuracy of 83% for license plate recognition, weapon detection, graffiti, and smoke detections. Although we achieved moderate accuracy in fire detection and lower performance in flood detection. Considering sociotechnical concerns, we evaluate CSIS’s ability to automate surveillance tasks. Contributions and Impact to IS Area: CSIS architecture introduces an interoperable architecture for intelligent surveillance, demonstrating a model of interoperability in real time;