基于Raspberry Pi的安全帽识别系统设计与实现  

Design and implementation of safety helmet recognition system based on Raspberry Pi

作　　者：王鑫[1] 史艳国[1] 李艳文[1] WANG Xin;SHI Yanguo;LI Yanwen(School of Mechanical Engineering,Yanshan University,Qinhuangdao,Hebei 066004,China)

机构地区：[1]燕山大学机械工程学院,河北秦皇岛066004

出　　处：《燕山大学学报》2024年第3期229-235,243,共8页Journal of Yanshan University

基　　金：河北省科学技术研究与发展计划资助项目(19221909D);教育部产学合作协同育人项目资助项目(202002072007);秦皇岛市科学技术研究与发展计划资助项目(201902A012)。

摘　　要：为了便于施工危险区域人员的自动化识别,提出了一种基于Raspberry Pi的安全帽识别算法。该算法将摄像头采集到的原始视频图像进行滤波、形态学等处理,再对图像中的安全帽进行识别。对于彩色安全帽,将原始图像转换到HSV空间,根据不同颜色色调阈值的设定同时识别红、黄、蓝三种颜色的安全帽,并结合形状特征剔除错误目标。对于白色安全帽,将原始图像转化成B通道下的灰度图像,解决了将黄色误检为白色的问题。采用V通道直方图均衡化的方法,提升了昏暗光线条件下的图像亮度。实验结果表明:在无需提前训练的情况下,算法对于单色安全帽识别准确率超过了91%,对于多色安全帽识别率超过了90%,为施工危险区域的安全隐患排查和作业管控提供了解决方案。In order to facilitate the automatic identification of personnel in construction hazardous areas,a algorithm based on Raspberry Pi to identification safety helmet is proposed.Filtering and morphology processing are applied to the original video image collected by the camera,and then identify the safety helmet in the image.For the color safety helmet,the original image is converted to the HSV space,and the three colors of red,yellow and blue safety helmets are simultaneously identified according to the setting of different color hue thresholds,and the wrong target is eliminated combined with the shape characteristics.For the white safety helmets,the original image is converted into a grayscale image under the B channel,which solves the problem of misdetecting yellow as white.The V-channel histogram equalization method is used to improve the image brightness under dim light conditions.The experimental results show that the algorithm was able to identify monochrome safety helmets with an accuracy of more than 91%and more than 90%recognition rate for color safety helmets without prior training,which provides a solution for the safety hazard investigation and operation control in construction dangerous areas.

关 键 词：Raspberry Pi 颜色识别 HSV空间 直方图均衡化 安全帽 

分 类 号：TP277[自动化与计算机技术—检测技术与自动化装置]