Research And Application Of Safety Helmet Detection And Identity Recognition System In J Logistics Park

In recent years,the safety management of logistics parks has become a hot issue.Due to the large number of vehicles entering and exiting the logistics park every day,and the intensive work of staff,accidents are highly likely to occur.However,full-time safety officers who work for a long time are prone to fatigue and high labor costs.Therefore,relying on full-time safety officers to supervise the operation personnel in the park cannot guarantee the safety of the park personnel.Based on the above pain points,this article proposes the research and application of a safety helmet detection and identity recognition system.This application is based on machine vision deep learning methods,which can effectively detect whether workers are wearing safety helmets and accurately identify their identity information,thus solving the security management problem of J logistics park.For the detection of safety helmets,this article uses the YOLOv7 object detection network algorithm.The dataset is taken from images captured in on-site monitoring videos and on-site photos,totaling 8500 pieces.In addition,in order to enhance the generalization ability of the model,this article designs a data processing method for safety helmets,which enhances and preprocesses images to achieve better recognition of safety helmets.For the identification of operators,this article adopts the Light CNN lightweight convolutional neural network algorithm,which can effectively recognize faces.Through experiments,it has been shown that this algorithm can meet real-time requirements and has high robustness.In unconstrained face recognition situations,the recognition rate can reach99.2%.The convolutional neural network based on Light CNN architecture is suitable for face datasets with high noise pollution.During the training and recognition process,this network extracts five important feature points in the face and corrects the face image based on the position of the feature points,thereby reducing the impact of image noise on recognition accuracy.This article combines the above two models for analysis.Firstly,using the YOLOv7 object detection model,head and helmet features are extracted to detect whether the helmet is worn or not,and facial information is extracted to support subsequent identity recognition.Secondly,in the face recognition process,the pre extracted facial information images are input into the Light CNN model for feature extraction to identify identity information.Finally,performance tests were conducted on the two models mentioned above.The helmet detection model was tested by capturing images from different lighting environments in on-site videos,achieving a recognition rate of 97.4%.The identity recognition model has been tested and compared in public datasets,with a recognition rate of 99.2%.The experimental results show that the recognition performance of the helmet detection algorithm and identity recognition algorithm used in the system is superior,verifying the feasibility and effectiveness of this method.After requirement analysis and background analysis,based on the algorithms of hard hat detection and identity recognition,this article has designed a comprehensive hard hat detection and identity recognition system for J Logistics Park.The system can achieve system management,system monitoring,and violation monitoring functions.After strict functional and performance testing,it has been confirmed that the system can meet the requirements of hard hat detection and identity recognition in J Logistics Park,Effectively solved the security management challenges of J Logistics Park.