Research On Spatial Pose Detection Method Of Roadheader Based On Roof Vision

In China's coal mining,the fully mechanized mining equipment with high automation level has greatly improved the mining efficiency,while the roadway excavation based on fully mechanized mining technology is inefficient,which leads to mining imbalance.It has become the industry consensus to solve the problem of mining imbalance by improving the automation level of boom roadheader.The automatic detection of the position and posture of boom roadheader is the first condition to realize its automatic cutting and improve its automation level.Machine vision is an important solution to realize the space pose autonomous detection of boom roadheader.Firstly,the paper analyzes the roof characteristics of the underground roadway environment,defines the position and posture of the roadheader and the corresponding coordinate system,so as to complete the construction of the mathematical model of the position and posture detection of the roadheader.Among them,the camera calibration technology based on plane target is studied,and the camera calibration experiment and calibration accuracy analysis are carried out.At the same time,the paper studies the coal mine roof image preprocessing method,pallet target extraction and feature detection method,realizes the segmentation of multiple pallet targets,completes the pallet target feature line detection and the calculation of straight line intersection,and then uses PNP algorithm to realize the solution of the static posture of roadheader.Then,the image matching technology and roadheader motion estimation method are studied to realize the recognition of the same tray target in adjacent image frames,and the dynamic pose of roadheader is solved.The pose optimization of roadheader in the process of motion is realized through nonlinear optimization function.Finally,the feasibility of this method is verified in the simulation tunnel.By comparing the position parameters measured by total station with the position parameters calculated by this method,the accuracy of position parameters is analyzed,and the correctness of attitude angle calculation is verified.