Electric power employee safety helmet wearing automatic identification method

Abstract

The invention provides an electric power employee safety helmet wearing automatic identification method, and relates to the technical field of mode identification and intelligent video analysis; a pedestrian detection model taking a pedestrian training sample as input is established, and pedestrian area detection of a test sample is realized through network parameter fine tuning; an SSD safety helmet detection model taking a safety helmet training sample set as input is established, a pedestrian upper body area image obtained by the pedestrian detection model is input into the trained SSD safety helmet detection model, and real-time and high-accuracy safety helmet wearing automatic identification is realized through SSD network parameter fine adjustment. According to the electric power employee helmet wearing automatic identification method, the efficient machine learning algorithm is utilized to directly detect the target, the detection precision is prevented from being influenced bycolor and edge feature changes caused by external environment changes, and the detection precision and the detection efficiency are improved by adopting two detection model cascade application modes at the same time.

Description

technical field [0001] The invention relates to the technical field of pattern recognition and intelligent video analysis, in particular to an automatic identification method for electric power employees wearing helmets. Background technique [0002] The electric safety helmet is an important safety protection device in electric work. The electric safety helmet maintains the safety of the staff's head and avoids fatal injuries to electric workers such as electric shock and smashing injuries. At present, the work site of the power sector is usually installed with a safety production video monitoring system, which collects and stores video data through the camera, and then monitors the safety status of people or objects at the work site through manual viewing. Such a supervision process increases labor costs, and at the same time, some unnecessary safety accidents occur due to subjective factors such as fatigue and negligence of the video viewing staff. Therefore, it is neces...

AI Technical Summary

Problems solved by technology

[0009] The present invention aims at the defects of the above-mentioned automatic identification method for helmet wearing in the existing electric power work site monitoring video, and provides an automatic identification method for electric power employee helmet wearing, which uses an efficient machine learning algorithm to directly detect the target, avoiding external Environmental changes lead to changes in color and edge features and affect detection accuracy. At the same time, two detection models are cascaded to improve detection accuracy and efficiency.

Images