Research On Recognition Of Tile Gap In Robot Paving Tiles

With the development of the construction industry,ceramic tile is widely used in the decoration of the floor and wall in the construction industry.In the construction process,the labor intensity is high and the construction efficiency is low when workers lay and paste ceramic tiles,at the same time,the labor cost of enterprises is rising continuously.The introduction of robot technology into the building decoration industry has become the key to solve such problems.When the robot lays ceramic tile,it needs to adjust the position and pose of ceramic tile according to the vision,in order to make the precision of ceramic tile laying meet the requirements.In this paper,a set of machine vision system was designed for brick seam recognition and pose adjustment,including selection of brick seam recognition scheme,establishment of image acquisition system,image processing of brick seam pictures,calculation of pose adjustment amount,construction of simulation system and simulation of robot tile laying.It mainly includes the following contents:(1)Briefly described the structure,size,quality and performance parameters of each part of the robot,ROS and Gazebo were used to build a tile-laying robot model based on actual parameters,and a tile-laying simulation environment was built.Added cameras,end effector suction cups,lighting sources and other related plug-ins to the virtual robot,used the Moveit!software package to control the motion of the robot arm,and use the joint controller to imitate the motor and reducer to control the rotation of the robot arm’s joints,so that the virtual robot can tile rooms just like real robots.(2)Introduced 4 mainstream image collection schemes,analyzed the pros and cons of each brick joint image collection scheme,and choose a machine vision scheme for brick joint image collection.The geometric model of brick seam imaging was established,the principle of tile pose adjustment was expounded,and the amount of pose adjustment was calculated.By considering the viewfinder range,object distance,spatial resolution and other factors,the appropriate hardware of the vision system was selected.(3)Introduced in detail the image processing flow of brick joints,using OpenCV to process brick joint images,including image gray-scale processing,median filtering,Canny edge detection and Hough line fitting.Aiming at the characteristics of the gray image and gradient map of the tile joint image,the dual-threshold Otsu algorithm was improved,which was used to select the high and low thresholds of the Canny operator,so that the Canny operator can extract the edge of the tile more efficiently.When the robot was laying the first tile,the preadjustment step was added to make the image processing result more accurate.A self-developed image processing flow for the joints of bricks was used to process the joints of bricks with different textures,which proved that the algorithm has strong applicability.(4)A simulation experiment was carried out according to the robot tile-laying process,which verified the correctness and applicability of image acquisition,image processing and pose adjustment in the tile-laying robot system,obtained the time of the robot paving tile adjustment stage and the tile paving accuracy after adjustment.