采煤机作业区域人员精确检测方法研究  

Research on precise detection method of personnel in shearer operation area

作　　者：魏东 王忠宾[1,2] 司垒[1] 谭超[1] 路绪良 WEI Dong;WANG Zhongbin;SI Lei;TAN Chao;LU Xuliang(School of Mechatronic Engineering,China University of Mining and Technology,Xuzhou 221116,China;Jiangsu Key Laboratory of Mine Mechanical and Electrical Equipment,China University of Mining and Technology,Xuzhou 221116,China)

机构地区：[1]中国矿业大学机电工程学院,江苏徐州221116 [2]中国矿业大学江苏省矿山机电装备高校重点实验室,江苏徐州221116

出　　处：《工矿自动化》2022年第2期19-28,共10页Journal Of Mine Automation

基　　金：国家重点研发计划资助项目(2020YFB1314200);国家自然科学基金青年基金资助项目(52074271);中国博士后科学基金特别资助项目(2020T130696)。

摘　　要：当前智能化采煤机已具有三维定位、记忆截割和远程监控等功能,但缺少采煤机作业区域误入人员的检测和预警保护功能,人员精确检测是亟待解决的关键问题之一。受综采工作面低照度、工况环境复杂多变影响,基于激光、射频、超声波等传感器的煤矿机电装备主动防撞预警技术应用受限,基于可见光传感器的防撞技术难以满足准确性和稳定性要求。搭建了基于红外热成像技术的采煤机作业区域人员精确检测系统架构,进而提出了人员精确检测方法:针对综采工作面红外图像噪声的高强度、不均匀特点,采用基于高斯掩码改进的多级导向滤波模型有效滤除红外图像噪声,并保留边缘信息;基于Lucas-Kanade光流法提取动态背景下的移动前景目标运动信息;采用基于图像局部信息权重的直觉模糊C均值聚类算法对采煤机作业区域红外图像信息进行分割,获取移动目标位置信息;基于形态学加权投票法对移动目标运动信息提取结果和红外图像信息分割结果进行融合,实现采煤机作业区域人员精确检测。在耿村矿21208综采工作面进行井下工业性试验,结果表明采煤机作业区域人员精确检测方法对现场人员的跟踪偏差平均值为0.1065像素,重叠比平均值为96.10%,平均单次处理时间为0.4908 s,满足现场应用需求。The current intelligent shearer has the functions of three-dimensional positioning,memory cutting and remote monitoring,but it lacks the detection and early warning protection function of personnel entering the shearer operation area by mistake.Therefore,the precise detection of personnel is one of the key problems to be solved urgently.Affected by the low illumination and complex working conditions of fully mechanized working face,the application of active anti-collision warning technology for coal mine electromechanical equipment based on laser,radio frequency,ultrasonic and other sensors is limited,and the anti-collision technology based on visible light sensor cannot meet the requirements of accuracy and stability.The system architecture of precise detection of personnel in shearer operation area based on infrared thermal imaging technology is built,and then the precise detection method of personnel is proposed.Aiming at the high intensity and uneven characteristics of the infrared image noise in fully mechanized working face,an improved multi-layer guided filter model based on Gauss mask-code is used to filter out infrared image noise effectively and retain the edge information.The moving foreground target motion information under dynamic background is extracted by optical flow method that based on Lucas-Kanade.The intuitionistic fuzzy C-means clustering algorithm based on the weight of local image information is used to segment the infrared image information of the shearer operation area so as to obtain the position information of the moving target.Based on the morphological weighted voting method,the extraction results of the moving target motion information and the infrared image information segmentation results are fused to realize the precise detection of personnel in the shearer operation area.The underground industrial test is carried out in 21208 fully mechanized working face of Gengcun Coal Mine.The results show that the average tracking deviation of the precise detection method for personnel in the

关 键 词：综采智能化 采煤机 人员精确检测 红外热成像 光流法 运动信息提取 图像分割 图像形态学 

分 类 号：TD421[矿业工程—矿山机电]