论文部分内容阅读
为解决行为安全领域不安全因素识别和交互分析困难的问题,构建基于深度学习的不安全因素识别和交互分析模型。首先,从“人-机-环”3方面构建不安全因素识别层,分别采用不同的深度学习结构识别作业人员行为属性、工作环境场景和操作设备工作状态的不安全因素;然后,通过因素交互层,采用关联和回归多值算法完成对不安全因素的交互分析;最后,通过输出显示层实现分析结果的表征。以某煤矿综采、掘进、通风3个生产活动类别的视频音频数据为例,通过Matlab操作平台选取最优深度学习结构,进行模型交互分析。结果表明,用该模型能实现对采煤面空顶作业、喷浆机故障仍然加料、主要通风机异常响动未停机检查等不安全因素的识别和交互分析,完成不安全行为的描述以及风险分级、行为痕迹的分类。
In order to solve the problem of identification and interaction analysis of unsafe factors in the field of behavioral safety, an insensitive factor identification and interaction analysis model based on deep learning is constructed. First of all, construct the recognition layer of insecurity factors from 3 aspects: human-machine-loop, and use different deep learning structures to identify the behavior attributes of workers, the working environment scenarios and the unsafe factors of the working conditions of the operating devices. Then, Factor interaction layer, the correlation and regression multivalue algorithm is used to complete the interaction analysis of the insecurity factors; finally, the output display layer is used to realize the characterization of the analysis result. Taking the video and audio data of three production activity categories of a fully mechanized mining, tunneling and ventilation in a coal mine as an example, the optimal deep learning structure is selected through the Matlab operating platform to analyze the model interaction. The results show that this model can realize the identification and interaction analysis of the unsafe factors such as the coal mining face empty operation, the shotcrete machine failure feeding, the main ventilator abnormal noises checking, the description of the unsafe behavior and the risk classification , The classification of behavior traces.