Method for detecting wearing state of safety helmet in dangerous working area in power field
A technology of working area and detection method, applied in the field of intelligent recognition, can solve problems such as large difference in target size and occlusion by helmets, and achieve the effect of strong adaptability and high efficiency
Active Publication Date: 2019-07-30
SHANDONG UNIV +2
View PDF9 Cites 15 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
For data training, more than 7,000 pictures have been marked, including various complex environments such as construction sites, power stations, electric towers, and indoors. Hard samples for training
Method used
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment
[0054] Such as figure 1 shown.
[0055] A method for detecting the wearing state of a helmet in a dangerous work area in the electric power field, the detection method comprising the following steps:
[0056] S1: Preprocess the dataset of hard hats; construct difficult samples on this basis, complete the labeling of specific scenes, and train on the improved tiny-yolo model to complete the pairing: people, people wearing hard hats, people without wearing hard hats The person in the hard hat, the classification of the above three types of situations;
[0057] S2: Perform frame extraction processing on the video captured by the mobile terminal, and obtain a deep representation of the extracted image through the neural network model;
[0058] S3: Complete the classification of features through the detection network, realize the above three types of detection: people, people wearing helmets, and people not wearing helmets, and complete the accelerated optimization of the mobile ...
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More PUM
Login to View More Abstract
The invention discloses a method for detecting the wearing state of a safety helmet in a dangerous working area in the power field. The method comprises the steps of S1, preprocessing a data set of the safety helmet; on this basis, constructing a difficult sample to complete the marking of a specific scene, carrying out training on the improved tiny-yolo model, and completing classification of thethree kinds of conditions of people wearing the safety helmets, people not wearing the safety helmets and people wearing the safety helmets; S2, performing frame extraction processing on the video captured by the mobile terminal, and obtaining deep representation of an extracted picture through a neural network model; S3, completing classification of the features through a detection network, realizing three kinds of detection of persons wearing safety helmets and persons not wearing safety helmets, and completing acceleration optimization of the mobile terminal on a TensorFlow Lite learningframework; and S4, performing a non-maximum suppression operation on all the detected bounds to filter out redundant boundary frames, thereby realizing classification detection of the target.
Description
technical field [0001] The invention relates to a method for detecting the wearing state of a helmet in a dangerous working area in the electric power field, and belongs to the technical field of intelligent identification. Background technique [0002] Substations, tower operations, and inspections of transmission lines have complex environments, and there are many factors that threaten people's lives. As the most critical part of the human body, the protection of the head is particularly important. Therefore, workers are required to wear safety helmets in such work scenarios. In recent years, many works have been carried out to try to solve the problem of hard hat wearing detection. The traditional safety helmet detection method is complicated in process, large in calculation, and has many false positives and false positives. With the development of artificial intelligence, many researchers have begun to combine deep learning technology with the problem of helmet detecti...
Claims
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More Application Information
Patent Timeline
Login to View More IPC IPC(8): G06K9/00G06K9/62
CPCG06V20/41G06F18/241G06F18/214
Inventor 聂礼强尹建华王英龙战新刚姚一杨朱建飞
Owner SHANDONG UNIV