基于深度学习的安全帽检测方法研究

Research on detection method of safety helmet based on deep learning

在电厂生产过程中,安全帽对于保障员工的安全具有非常重要的作用。工作人员在进行生产操作时未正确穿戴安全帽有可能直接导致事故的发生。在开发电厂不安全行为视频检测系统中,安全帽的检测将是一项需要解决的关键问题。本文通过收集现场图片信息和人工标注的方法,构建了训练集和测试集。通过采用深度学习算法,在数据集上得到了一种具备安全帽检测的神经网络模型。经验证,该模型在构建的测试集上达到了良好的检测效果。

Helmet plays a very important role in ensuring the safety of employees in the production process of power plants.Improper wearing of safety helmet during production operation may cause accidents directly.In the development of video detection system for unsafe behaviors in power plants,the detection of safety helmets will be a key problem to be solved.In this paper,the training set and the test set are constructed by collecting on-site picture information and manually tagging.A neural network model with helmet detection is obtained on the data set by using the deep learning algorithm.It is verified that the model has achieved good detection results on the constructed test set.