Warehouse video monitoring system based on edge computing

Abstract

The invention discloses a warehouse video monitoring system based on edge computing. The warehouse video monitoring system comprises a monitoring acquisition terminal module, an edge node module, a peripheral network transmission module and a terminal control end display module which are connected in sequence, and the system also comprises a power supply module supplying power to each module. By migrating the analyzed and processed task to the edge node module, on the premise of ensuring the data reliability, a large amount of real-time edge data generated by the monitoring acquisition terminal module in a short time is preprocessed, so that the video analysis speed is improved, the processing and transmission time delay is reduced, and the real-time performance of the monitoring video stream is ensured. Meanwhile, an improved two-frame difference method and projection method fused algorithm is applied to detect the moving object, and the moving object is stored only when the moving object is detected in the monitoring picture, so that the credibility of evidence information is enhanced, the storage space utilization rate of video data is improved, and a large amount of storage space is saved.

Description

technical field [0001] The invention relates to the technical field of intelligent video monitoring, in particular to a warehouse video monitoring system based on edge computing. Background technique [0002] Video surveillance system is mainly used in video processing, target query and personnel tracking, etc., and has gradually become an important guarantee for urban public safety. The video resolution collected by the front-end camera of the traditional video surveillance system is high, and the amount of video data is large. The calculation and transmission bandwidth load of the traditional cloud mode video surveillance system is heavy; the video processing capability of the existing intelligent surveillance system is insufficient; the existing camera There are insufficient computing power, delay and bandwidth of uploading data, resulting in problems such as large missed detection of target information and low detection efficiency. [0003] In view of the unreliable det...

AI Technical Summary

Problems solved by technology

The video resolution collected by the front-end camera of the traditional video surveillance system is relatively high, and the amount of video data is large. The calculation and transmission bandwidth load of the traditional cloud mode video surveillance system is relatively heavy; the video processing capability of the existing intelligent surveillance system is insufficient; the existing camera Insufficient computing power, delay and bandwidth of uploading data, etc., resulting in large missed detection of target information and low detection efficiency

[0003] In view of the unreliable detection rate of moving objects proposed above and the time-consuming and space-consuming processing of data, and the inability to guarantee real-time performance, there are currently two algorithms:

[0004] ⑴The weighted behavior recognition algorithm reduces the amount of calculation in video analysis and processing, and solves the problem of high energy consumption, but this method only reduces the amount of calculation in the behavior analysis stage, and does not process redundant video frames

[0005] (2) The dynamic task scheduling algorithm migrates part of the computing tasks to the cloud to complete. However, in scenarios with large traffic such as stations and airports, the amount of data increases sharply in a short period of time, and the target recognition rate decreases.

[0006] Although the above method has improved the current situation that the cloud platform monitoring system has high requirements for bandwidth and computing resources, it has not fundamentally solved the problem, that is, the video frame is not filtered, and too much redundant edge data is transmitted, which makes it difficult to migrate computing tasks no matter what. The total amount of calculation has not been reduced, and the storage space is in short supply

Images