Design And Implementation Of Detection System Of Wearing Helmets Based On Deep Learning

Nowadays,production safety is more and more valued by enterprises.But unlike the highly modernized production level of enterprises,security protection measures are still very backward.For example,the supervision work of helmet wear almost achieve by people,so the supervision cost is high and the efficiency is low.In recent years,deep learning technology has a great-leap-forward development.So some researchers have started research on helmet detection based on deep learning.However,most of them use the color feature of the helmet to detect.Because interfere with the background color,detection accuracy rate is low.According to the above problems,this paper designs and implements a set of intelligent safety protection system based on deep learning.The main function of the system is to perform real-time security warning for the camera monitoring area,and realize the functions of pedestrian detection,helmet wearing detection,intelligent alarm and system management.This paper uses the SSD(Single Shot MultiBox Detector)module to train the pedestrian detection model and the helmet detection model to to realize the relevant detection functions.And this paper mainly describes the design and implementation of the system from the angle of helmet detection.Due to the slow speed and low precision of the traditional helmet wearing detection method,the system helmet detection module designed and implemented the helmet detection in the pedestrian area.Firstly,the pedestrian detection model is used to detect the pedestrians in the video,and then the detected pedestrian area is transmitted to the helmet detection module to perform the helmet detection on the pedestrian area.This detection method effectively improves the detection accuracy of the wearing of the helmet,and effectively reduces the detection time,making the system more reliable and real-time.Finally,the system is connected with the monitoring equipment to achieve the detection of wearing a helmet by pedestrians within the monitoring range.This paper mainly completes the following work:1.Under the Tensorflow framework,use the SSD algorithm to train the pedestrian detection model and the helmet detection model,including acquire data set,data enhancement and labeling data set,training models and test models.2.According to the actual needs,this paper designs for the intelligent safety protection system in detail,and describes the requirements analysis,architecture design,functional module design,business process design and database design of the system.3.The specific implementation of the intelligent safety protection system.According to the design of the system to implement the functional modules,the detection model is integrated into the system to achieve the detection of pedestrians and helmets,and finally complete the connecting with the monitoring equipment and test.Through the test,the system can alarm pedestrians and non-wearing helmets in the alarm area according to the set alarm parameters,which has high accuracy rate and achieves the expected effect,which can meet the actual supervision requirements.