Research On Model Construction Of Crane Robot And Vision Tracking For Hook Trajectory

The safe and reliable operation of the overhead travelling crane is essential to the production and transportation of materials for company.The experimental platform of crane robot system was designed and built in the laboratory to study the running characteristics of the overhead traveling crane and the obstacle avoidance and monitoring of the visual system in order to improve the automatic and intelligent level.The results of this paper can be used for reference in improving the automation and intelligence of crane robot.The main research contents are as follows: This paper analyzes the principle,structure and design the executive system and control system of the crane robot and select the visual sensor.The kinematic coordinate system of the crane robot and the simplified coordinate system are established and analyzed.The model of overhead crane robot is established.In view of the different movement combinations of the main car,the trolley and the hook from the main parts of the crane,the trajectory model of the geometric center of gravity of the hanger is established and analyzed.The track of the center of gravity of the stationary or linear motion of the hook and the trajectory when the hook is in a plane swing is analyzed,based on the robot motion state of the flexible cable crane hook.The binocular stereoscopic vision model of the crane is established.The relationship between the camera small hole imaging model,the camera parameter model and the barycenter imaging coordinate of the hook and the linear motion speed of the crane robot are analyzed and the parallel binocular vision model of the crane is established.RGB different color threshold selection method is used to segment the target image of the hook.Cut the hook motion sequence image,and then use the color-based image segmentation method to separate the hook from the complex background.Based on the SIFT feature point matching method,the image matching of hook target is realized,and the H-S optical flow method is used to study vision recognition and tracking.