AI-Driven Real-Time Surveillance for Dynamic Activity Monitoring

In all sectors and applications, surveillance is essential for security and safety. Continuous recording causes storage issues for traditional systems. This study presents an intelligent surveillance system that uses AI and computer vision to selectively record when a human face is identified. The Haar Cascade Classifier with Python and OpenCVis usedfor real-time face detection. According to experimental results, our solution outperforms traditional methods by reducing storage by 50% while retaining an average detection accuracy of 85%. We also contrast the performance of our system with other AI-based surveillance techniques like YOLO and Mobile-net, emphasizing the trade-offs between computing efficiency and detection speed.