摘要
为了实现掘进机位姿参数的实时检测,构建了融合机器视觉和惯性导航的掘进机位姿参数检测系统,利用惯性导航系统检测掘进机姿态参数并融合机器视觉检测掘进机位置参数。为了克服粉尘和光线带来的影响,该系统配套了相机自清洁仪和补光光源。针对矿井现场环境,为减少人为干预,以巷道顶板托盘为视觉特征进行掘进机位置检测。针对视觉定位问题,研究了图像目标和位置检测的方法。为了实现掘进时间序列的位姿检测,采用机器视觉和惯性导航信息融合的方法实现连续帧间的目标匹配,得到连续帧间掘进机位姿参数的变化量;结合掘进机初始的位姿参数,解算掘进机的实时位姿。为了验证算法的可靠性和检测精度,搭建了模拟巷道进行实验。结果表明:该系统能够实现掘进机连续帧间的目标跟踪,在掘进机进尺为100 m时,X轴方向定位误差为17.8 mm,Y轴方向定位误差为74.7 mm,Z轴定位误差为50.3 mm,满足断面成形精度要求,为实现掘进机工作面自动化、智能化提供了助力。
To achieve the real-time detection of position and orientation parameters of the roadheader,a roadheader position and orientation detection system with fusion of machine vision and inertial navigation was constructed.The inertial navigation system was used to detect the orientation parameters of the roadheader and integrated machine vision to detect the position parameters of the roadheader.To overcome the impact of dust and light,this system is equipped with a camera self-cleaning device and supplementary lighting source.Aiming at the mine site environment and to reduce human intervention,took the roadway roof pallet as the visual features for the detection of the position of the roadheader.Aiming at the problem of visual positioning,the methods of image target and position detection were researched.In order to realize the position and orientation detection of the roadheader time series,the method of machine vision and inertial navigation information fusion was used to realize the target matching between consecutive frames,and the change of the roadheader position and orientation parameters between consecutive frames were obtained.Combined with the initial position and orientation parameters of the roadheader,the real-time position and orientation of the roadheader was computed.To verify the reliability of the algorithm and the detection precision,the simulation roadway was built to do experiment.The results show that this system can achieve thetarget tracking between consecutive frames.The positioning error in the X-axis direction is 17.8 mm,that in the Y-axis direction is 74.7 mm and that in the Z-axis direction is 50.3 mm when the progress distance of roadheader is 100 m,which meets the requirements for cross-section forming accuracy and provides a help for realizing automatic and intelligence of the excavation face.
作者
贾曲
Jia Qu(Shanxi Tiandi Coal Mining Machinery Co.,Ltd.,Taiyuan 030006,China;CCTEG Taiyuan Research Institute,Co.,Ltd.,Taiyuan 030006,China;China National Engineering Laboratory for Coal Mining Machinery,Taiyuan 030006,China)
出处
《煤矿机械》
2025年第5期209-214,共6页
Coal Mine Machinery
基金
山西省重点研发计划项目(202202020101005)。
关键词
掘进机
托盘
机器视觉
惯性导航
多信息融合
位姿检测
roadheader
pallet
machine vision
inertial navigation
multi-information fusion
position and orientation detection
