基于自适应背景差分与深度学习的矿山巷道不安全行为的自动识别  

Automatic Detection of Unsafe Acts in Mine Roadway Based on Adaptive Background Subtraction and Deep Learning

作　　者：韩苗 许可 伍书缘 王汉生[3] Miao Han;Ke Xu;Shuyuan Wu;Hansheng Wang(School of Mathematics,China University of Mining and Technology;School of Statistics,University of International Business and Economics;Guanghua School of Management,Peking University)

机构地区：[1]中国矿业大学数学学院 [2]对外经济贸易大学统计学院 [3]北京大学光华管理学院

出　　处：《经济管理学刊》2023年第2期75-96,共22页Quarterly Journal of Economics and Management

基　　金：国家自然科学基金项目(12271012,11831008,12001102);统计与数据科学高等理论与应用重点实验室开放研究基金(KLATASDS⁃MOE⁃ECNU⁃KLATASDS2101);对外经济贸易大学中央高校基本科研业务费专项资金(CXTD13⁃04)对本文研究的支持。

摘　　要：煤炭开采行业是公认的高危行业,其中人的行为因素是造成绝大多数事故的直接原因。及时提醒、纠正矿工的不安全行为是避免煤矿事故最重要且最有效的方法。本文基于自适应背景差分模型,提出了一个三阶段优化算法,通过异常值发现、异常值区域像素点平面平滑以及连通域分析算法,实现了矿工子图数据集的有效提取。特别是在异常值发现中,本文采用中位数估计方法,通过分布式中位数计算,求得标准差更稳健的中位数估计量,替代传统矩估计量构造阈值,获得更加稳健的异常值区域,进而提取高质量的矿工子图数据集。实验结果表明,与传统估计方法相比,基于中位数估计得到的标准差所构造的阈值,提取的矿工子图准确率更高,提取的结果更稳健。同时,本文直接对矿工子图标注类别,避免进行边界框的标注,操作简单,快捷高效。最后,本文使用MobileNet模型进行迁移学习,结果进一步表明,中位数估计方法得到的矿工子图数据集的质量优于传统估计方法,子图分类准确度更理想,能够有效地识别矿工的不安全行为。China's coal-based energy resources endowment and the current stage of its economic and social development determine that the economic and social development will remain inseparable from coal for a considerable period of time in the future.Even with the“dual carbon”target,coal still plays its role as a basic energy source,and provides energy support for cconomic and social de-velopment.The coal mining industry is recognized as a high-risk industry where human behavioral factors are the direct cause of the vast majority of accidents.The monitoring system enables timely detection of unsafe miner bchavior and timely intervention or change of unsafe miner bchavior,which can effec-tively prevent accidents from occurring.However,safety and security in coal mines can be compro-mised by the responsibility,work fatigue and efficiency of safety monitoring staff.Therefore,it is important to study automation methods based on machine leamning to replace manual monitoring using artificial intelligence technology in order to achieve automatic identification of unsafe behavior in underground mines in a safe and efficient manner over a long period of time to ensure coal mine safety.Specifically,it is hoped that the high-frequency,high-resolution image data captured by survillance cameras will be used as input to develop appropriate statistical and deep learming models for the timely detection and correction of unsafe bechaviors in mine production operations.Taking the identi-fication of ilgal bchaviors by miners on monkey vehicles as an example,the essence of the problem is target detection and target identification.The current mainstream algorithms can be divided into two-stage methods and onc-stage methods.Two-stage methods,such as R-CNN and Faster R-CNN series algorithms based on candidate regions.One-stage methods,such as Yolo and SSD algorithms,which are currently receiving great attention.Although these methods have played an important role in target detection,they are difficult to apply to the mine safety management is

关 键 词：不安全行为 背景差分 分布式中位数计算 迁移学习 

分 类 号：O213[理学—概率论与数理统计]