基于机器视觉掘进机位姿检测应用现状与趋势

Application status and trend of roadheader pose detection based on machine vision

基于机器视觉技术利用防爆相机获取巷道环境信息,安装方便且具有非接触、获取信息丰富的特点,在煤矿智能化进程中得到了广泛应用。在煤炭开采中,掘进机作为巷道掘进的主要装备,具有截割、转载、运输的功能,且掘进机的操作主要依赖人工操作,制约了巷道掘进的效率,而掘进机的位姿检测不仅为工人操作提供了参考,而且有利于掘进机的自动化以及智能化操作,进而有利于煤矿智能化进程的推进,不仅能缓解采掘失衡的矛盾,同时能降低工人的劳动强度。本文分析了近几年机器视觉技术在悬臂式掘进机位姿检测中的应用现状与发展趋势,研究了悬臂式掘进机各种位姿检测方法中所涉及的图像预处理以及位姿检测模型的共性问题,提出了机器视觉技术在掘进机位姿检测的应用中,图像采集硬件、位姿检测的冗余性以及特征提取的精确性方面仍需要进一步攻关与突破。

Relying on the characteristics of non-contact and rich information acquisition,machine vision technology has been widely used in the process of intelligent coal mining.In coal mining,the roadheader,as the main equipment of roadway tunneling,has the functions of cutting,transferring and transportation.Moreover,the operation of the roadheader mainly depends on manual operation,which restricts the efficiency of roadway tunneling.The position and posture detection of the roadheader not only provides reference for workers operation,but also is conducive to the automation and intelligent operation of the roadheader,which is also conducive to the advancement of the intelligent process of the coal mine.It can not only alleviate the contradiction of mining imbalance,and reduce the labor intensity of workers.This paper analyzes the application status and development trend of machine vision technology in the position and posture detection of cantilever roadheader in recent years,studies the common problems of image preprocessing and position and posture detection model involved in various position and posture detection methods of cantilever roadheader.The redundancy of pose detection and the accuracy of feature extraction still need further research and breakthrough.