Video analysis method and system for illegal behaviors of safe operation

Abstract

The invention discloses a video analysis system and method for illegal behaviors of safe operation, and belongs to the field of image processing. The system comprises a hardware system and a software system. The software system comprises a video analysis subsystem which runs on the hardware platform of the video analysis workstation; and the video analysis subsystem comprises a real-time video reading module, a video analysis module, an illegal behavior picture storage module, a network alarm module and a log recording module. The video analysis subsystem can analyze videos collected by a cigarette factory roller operation camera, and automatic detection of four typical illegal behaviors that employees do not wear safety helmets and anti-skid shoes, roller switches are not padlocked and signboards are not hung is achieved; a video analysis method for detecting the four illegal behaviors is designed on the basis of a traditional visual detection algorithm such as safety helmet detection; the test result verifies the correctness of the system function; and according to the method, the capability of preventing the safety risk of roller operation is improved.

Description

technical field [0001] The invention relates to the field of computer vision and image processing and the field of drum operation safety, in particular to a video analysis method and system for safety operation violations. Background technique [0002] In some dangerous places, such as during the operation of the drum in the tobacco factory, employees do not wear safety helmets, non-slip shoes, padlock the drum switch, and do not hang signs. Sometimes when entering the drum, the alarm sign will not be set up and locked, which will have potential safety hazards for future safe production. The legal work process is to check the wear of the workers before entering the drum, including safety helmets and non-slip shoes, and then check the locks and warning signs. When the lock and sign are hung on the specific drum knob switch, Only then is it allowed to enter the drum for work. [0003] Traditionally, manual supervision by security personnel is time-consuming and labor-intensi...

AI Technical Summary

Problems solved by technology

[0003] Traditionally, it is manually supervised by security personnel, which is time-consuming and labor-intensive, and is prone to human negligence. Therefore, the industry has begun to develop automatic detection

However, in the artificial intelligence detection system, the detection will inevitably lead to false detection and false alarm.

Moreover, the problem of false detection is a common problem faced by existing artificial intelligence detection systems, especially in complex environments, which will directly affect the function of the system

[0004] Therefore, the inventor proposed a target monitoring algorithm based on deep learning to overcome the problems of low precision and poor resolution in safety monitoring for the safety detection system before dangerous operations by enterprise construction personnel.

Images