Large-scale distributed monitoring video data processing method and device

Abstract

The invention provides a large-scale distributed monitoring video data processing method and device. The method comprises the following steps of at a video data analysis front end, obtaining monitoring video stream data, carrying out dynamic modeling on the video stream data, extracting a target area in the video stream data, calculating a feature value of the target area, extracting target metadata data information based on the feature value through target feature identification, and generating a video data processing result; and at a video data processing rear end, obtaining data processing result metadata messages and pictures, and carrying out secondary processing on the data processing result metadata messages and pictures. According to the method and the device, through utilization of a high-performance camera device deployed at a front end, the high-definition monitoring video data is analyzed and processed at the front end; compared with the mode of analyzing the high-definition video data at a rear end, the mode has the advantage of reducing network transmission bandwidths and video analysis and computing resources; and the new high-definition video data analysis and transmission solution scheme for large-scale high-definition monitoring video deployment and service application is provided.

Description

technical field [0001] The invention relates to the technical field of data processing, in particular to a large-scale distributed surveillance video data processing method and device. Background technique [0002] With the gradual development of surveillance video systems from analog and SD to HD, UHD and 4K, surveillance video data analysis in many fields such as smart cities and smart transportation is the most basic requirement for the application of surveillance video systems. Because the video data analysis of the traditional high-definition surveillance video system mostly relies on the back-end large PC server for processing, it is necessary to transmit the high-definition video data to the server room through the network for analysis and calculation, and then store the target analysis results, especially when the high-definition surveillance video reaches When thousands or even tens of thousands of channels and data reach PB level or above, the cost of surveillance ...

AI Technical Summary

Problems solved by technology

[0004] The technical problem to be solved by the present invention is to provide a method for realizing high-definition video data analysis relying on a large-scale distributed front-end high-performance camera device to solve the existing high-definition surveillance video system. Network transmission of video in China, data analysis and computing resources become bottlenecks

Images