Method and device for monitoring safety helmet wearing of workers

Abstract

The invention provides a method and device for monitoring the safety helmet wearing of workers. The method includes the following steps that: a video file collected by a camera is acquired; the videofile is converted into RGB images frame by frame; the coordinates of human body nasal bones in the RGB images are obtained through a trained two-branch multi-level convolutional neural network; the RGB images are sequentially inputted into a trained Mask-RCNN, and instance segmentation is performed on safety helmets in the RGB images, so that the binary masks of safety helmet parts in the corresponding RGB images, which are outputted by the trained Mask-RCNN are obtained; and the centroid coordinates of the safety helmets in the RGB images are calculated through a preset first formula and a preset second formula according to the binary masks of the safety helmet parts in the RGB images; Euclidean distances between the human body nasal bones and the safety helmets are calculated through a preset third formula according to the coordinates of the human body nasal bones and the centroid coordinates of the safety helmets; and whether the Euclidean distances are larger than a preset distancethreshold is judged, and if the Euclidean distances are larger than the preset distance threshold, the workers do not wear the safety helmets correctly when working, if the Euclidean distances are not larger than the preset distance threshold, the workers wear the safety helmets correctly when working.

Description

technical field [0001] The invention relates to the technical field of machine vision, in particular to a method and device for monitoring the wearing of safety helmets by workers. Background technique [0002] At present, safety issues for migrant workers occur frequently. The traditional manual monitoring method and the responsibility of supervisors are the key to such safety issues. In addition, false positives and false negatives of manual monitoring are also inherent flaws. The cost of manpower is also an important factor considered by existing companies. [0003] Due to the low safety awareness of the staff, or due to some external factors, the staff forgot to wear the helmet, resulting in frequent safety accidents. It can be seen that the accurate detection of the helmet can remind the staff who do not wear the helmet to wear it. , which is beneficial to avoid safety accidents to a certain extent. [0004] In the prior art, the detection method for wearing safety h...

AI Technical Summary

Problems solved by technology

[0006] The present invention provides a method and device for monitoring workers wearing safety helmets, which solves the problem of manually judging whether to wear safety helmets from images due to the large monitoring area. Quality, on the other hand, is also dependent on the supervisor's sense of responsibility and fatigue, which can easily lead to inaccurate monitoring, and sometimes it is impossible to accurately judge whether users wear helmets, so that it is impossible to completely avoid safety accidents caused by helmets technical issues

Images