基于DBN-SVM的电力智慧工地异常行为识别

Abnormal detection of safety helmet in electric smart construction site based on DBN-SVM

随着电网的迅速发展,施工现场作业也出现了各种各样的情况,针对智慧工地背景下,施工人员安全防护用具的识别问题,提出基于DBN-SVM的电力智慧工地异常行为识别方法。通过在复杂场景中收集工作人员佩戴安全帽和不戴安全帽的图片,从而对数据图像预处理使其满足DBN模型的训练要求,然后搭建DBN-SVM模型,使用SVM方法进行分类识别,实现目标检测。实验结果表明这种方法可以有效地用于识别防护和安全设备,改善建筑工地工作人员的安全管理现状。

With the rapid development of the power grid,a variety of situations have also appeared in the construction site operation.In view of the identification problem of safety protection equipment for construction personnel under the background of smart site,this paper proposes a DBN-SVM based abnormal behavior identification method for the power smart site.Through collecting pictures of staff wearing and not wearing helmets in complex scenes,the data and images are preprocessed to meet the training requirements of DBN model.Then,the DBN-SVM model is built,and the SVM method is used for classification recognition and target detection.The experimental results show that this method can be effectively used to identify protection and safety equipment and improve the safety management status of construction site workers.