Intelligent Surveillance Camera System Based on Object Tracking

In an increasingly growing urban context, managing, monitoring and ensuring security is a big challenge for authorities and social organizations. Surveillance cameras have become a useful tool to monitor and analyze human activity and behavior in important areas such as streets, train stations, airports, hospitals, shopping centers and other areas. To solve this problem, many methods and algorithms have been proposed and developed. And wonderfully, the development of artificial intelligence and deep learning has also opened up new potentials in solving this problem. Therefore, an intelligent surveillance camera system based on deep learning has been proposed to be built in this paper. To process each frame individually, the system first extracts a series of frames from the input video. Next, a convolutional neural network (CNN) model is used to detect objects on each image. The YOLOv8 model is used to implement and develop the proposed system. This model is one of the most advanced human object detection models and has very high accuracy with an mAP50 value of 0.878. Besides, the Deep SORT model is used for object tracking. Then, the system checks whether the detected object is in the warning area or not. If the system detects people in the warning area and during the warning time, the system will issue a warning to the loudspeaker and display the warning status. At the same time, the system will save the object's movement process. Experimental results demonstrate the effectiveness of the proposed system.