Smoking behavior identification method and system, computer equipment and storage medium

Abstract

The invention discloses a smoking behavior identification method and system, computer equipment and a storage medium. According to the system, a camera device collects a current video stream in real time and sends the current video stream to an edge hardware device; the edge manager carries out resource scheduling on all the edge nodes and sends the determined target node identifier to the edge hardware equipment; the edge hardware equipment sends the current video stream to a target edge node corresponding to the target node identifier; the target edge node adopts a smoking behavior recognition model based on a YOLO network to perform smoking behavior recognition on the current video stream, obtains a smoking behavior recognition result, and sends the smoking behavior result to the monitoring device; the monitoring device receives the smoking behavior recognition result, the smoking behavior recognition result is displayed on a browser of the monitoring device, and alarm tracking processing is conducted on the target detection object with the smoking behavior recognition result being that the smoking behavior exists. The system can improve the accuracy and precision of smoking behavior recognition, and reduce the network bandwidth pressure and cost.

Description

technical field [0001] The invention relates to the technical field of artificial intelligence, in particular to a smoking behavior recognition method, system, computer equipment and storage medium. Background technique [0002] Illegal smoking in non-smoking places not only endangers one's own life safety, but also causes environmental pollution and even serious safety hazards. For example, illegal smoking in gas stations or other non-smoking places where fireworks are strictly prohibited. In the prior art, a smoking behavior recognition system is used to recognize real-time detected images of management personnel, and the recognized smoking behavior images are recorded and archived to provide evidence for accountability. [0003] Existing smoking behavior recognition systems generally use cloud servers to identify smoking behaviors from video streams collected by camera equipment. Due to the large amount of data transmitted by video streams, there are stricter requirements...

AI Technical Summary

Problems solved by technology

[0004] Embodiments of the present invention provide a smoking behavior recognition method, system, computer equipment, and storage medium to solve the problems of large delay, low accuracy, high network pressure and high cost in existing smoking behavior recognition

Images