基于智能视频图像分析的安全帽识别

为解决石油行业工作现场人员不佩戴安全帽的情况,避免因违规操作或违反规定,造成不必要的伤亡,提出一套基于施工现场安防监控系统的智能安全帽检测系统。设计一套综合利用安全帽的颜色、轮廓,以及多层神经网络分类建立的统计模型,对是否佩戴安全帽进行检测。结合实例,验证了算法的有效性和系统的可行性。实例结果表明,该系统能很好检测出现场没佩戴安全帽的人员,一定程度上杜绝了安全隐患,减少了事故的发生。

Safety helmet recognition based on intelligent video image analysis

To solve the situation that the workers in the petroleum industry do not wear helmets,and to avoid unnecessary operations resulting in unnecessary injuries,a smart helmet detection system based on the construction site security monitoring system to detect the presence of a worker without wearing a helmet was presented.A set of statistical models in which the color,contour,and multi-layer neural network classification of helmets were comprehensively utilized to detect workers’helmets was designed.The feasibility of the method was validated by practical application.The practical application shows that the proposed algorithm can well detect the presence of personnel without a safety helmet in the field,to a certain extent,eliminating hidden dangers and reducing the occurrence of accidents.