Research On Key Technology Of Mine Blasting Robot Based On 2D-3D Vision

With an increasingly aging population and rapid development of industrial automation in China,robots have become the current industrial development trend to replace human work.Among them,industrial robots have developed rapidly and are widely used in various fields such as industrial grasping,sorting,palletizing,spraying,etc.Which greatly reduces the amount of manual participation and greatly improves industrial production efficiency.The combination of 2D-3D vision and industrial robots is an emerging technology in the current industry.This paper takes industrial robots and special blasting vehicles as the research objects,and focuses on 2D-3D vision,robot system calibration,two-dimensional mine hole recognition and three-dimensional hole positioning,hole posture calculation,robotic arm trajectory planning and other technologies.2D-3D vision-based mine blasting robot autonomously fills explosives system.The main contents include:1.Robot system calibration: Robot system calibration includes RGB camera calibration,robot hand-eye calibration,and tool calibration,etc.RGB camera calibration mainly obtains the internal and external parameters of the camera.The conversion relationship from the camera coordinate system to the robot end flange coordinate system can be obtained by hand-eye calibration,In the process of hand-eye calibration,this thesis proposes the use of geometric methods to determine the conversion relationship from a point in space to the camera coordinate system.the conversion relationship from the tool to the robot end flange is completed by tool calibration.The ultimate goal of the robot system calibration is to convert the pixel coordinates of the object in each frame of image into the robot base coordinate system,and then control the robot end grabbing tool to move to the center of the blasthole.2.Hole recognition and location: This project uses YOLOV3's target detection algorithm,first trains the weight file according to its own hole data set,uses the target recognition algorithm YOLOV3 to detect the blast hole,and then uses the RGB camera calibration result to center the blast hole Pixel coordinates are converted to spatial coordinates in the camera coordinate system.3.Pose calculation and robot trajectory planning: First,according to the robot hand-eye calibration and the coordinates of the hole center in the camera coordinate system,study the position and posture of the hole center in the robot base coordinates,and then study the robot trajectory planning in the Cartesian coordinate system.That is,according to the current position of the robot's pose and the target position of the robot's pose,the robot's motion trajectory is planned.The content of this thesis is the calibration of the robot system,the recognition and positioning of the hole,the calculation of the pose of the hole and the trajectory planning of the robot.These contents are the key technologies of mine blasting robot system,and these key technologies have high practicability and accuracy in practical applications.