DEFORMATION CAMERA RECOGNITION SYSTEM USING NETWORK VIDEO TRANSMISSION PATTERN ANALYSIS BASED ON MACHINE LEARNING AND THE METHOD THEREOF

摘要

The present invention provides a modified camera recognition system using machine learning-based network video data transmission pattern analysis, which is capable of recognizing and blocking an illegal video transmission situation in real time. According to one embodiment of the present invention, the system comprises: an access point (AP) (500) relaying Internet access of modified cameras (400); a packet detection device (200) detecting a video data packet of the modified camera (400) outputted from the AP (500); an operation server (100) detecting a pattern matched with a predetermined pattern criterion from the data packet detected from the packet detection device (200), and controlling the packet detection device (200) to block the corresponding video data; and a manager terminal (300) transmitting a command for blocking or transmitting the corresponding video data to the operation server (100) when receiving detection information of the operation server (100). The pattern criterion includes at least one among an uplink retaining time, a protocol, and the size of the data packet.