Method and system for intelligent identification of network equipment

Abstract

The invention discloses a method for intelligent identification of network equipment. The method comprises: the network equipment receives the monitoring area designated by the server, and determines the corresponding video shooting device according to the designated monitoring area; command, and then obtain and decode the video source code; the network device extracts the video frame in the video stream, and determines the preset features contained in the video frame, and finally the network device identifies the elements in the video frame. The invention also discloses an intelligent identification system for network equipment. The present invention specifies a specific video recognition area through the server, and focuses on collecting video streams in the area, and calls main computing resources to perform video decoding and video recognition. The present invention can reduce unnecessary decoding times by reducing the number of decoding Video recognition overhead, using very small computing resources, quickly and effectively performs video recognition on network devices.

Description

technical field [0001] The invention relates to the field of video identification, in particular to a method and system for intelligent identification of network equipment. Background technique [0002] Video recognition mainly includes three links: front-end video information collection and transmission, intermediate video detection, and back-end analysis and processing. Video recognition requires front-end video capture and video capture equipment to provide clear and stable video signals, and the quality of video signals will directly affect the effect of video recognition. Then through the intelligent analysis module embedded in the middle, it can identify, detect and analyze the video picture, filter out interference, and mark the target and trajectory of the abnormal situation in the video picture. Among them, the intelligent video analysis module is an algorithm based on the principles of artificial intelligence and pattern recognition. [0003] The existing video r...

AI Technical Summary

Problems solved by technology

[0003] The existing video recognition processing method usually encodes the video stream and transmits it to the server through the network device for video inspection, analysis and recognition. Under the guidance of the edge computing idea, the work of video recognition can be used as a video acquisition terminal It is carried out directly on the smart video shooting equipment of the system. This solution has very high requirements on the video shooting equipment. A large number of existing traditional surveillance video shooting equipment cannot be competent in this system, resulting in great waste; The video code is transmitted to the intelligent network device, decoded on the intelligent network device, and then inspected, analyzed and identified. In this solution, the intelligent network device receives the video streams of all video acquisition devices within the control range for video identification, carrying a large number of The computing work requires high performance requirements for intelligent network devices. Due to the need to receive and decode a large number of video streams at the same time, the performance of intelligent network devices is a very obvious bottleneck in this system.

Images