Design Of Intelligent Sole Polishing System Based On Monocular Coded Structured Ligh

As footwear made in China accounts for a growing share of the global consumer market,the need for footwear manufacturers to increase production and productivity is growing by the day.In the shoe industry,soles and uppers are made separately and then glued together with shoe glue,which requires the inner edge of the sole to be roughened before gluing to improve the adhesion of the shoe glue.At the present stage,the sole grinding work is mainly done manually by workers using grinding machines,which is inefficient and the dust generated by grinding affects workers’ health.The reliance on manual grinding has become one of the major issues limiting the efficiency of shoe manufacturing.To solve this dilemma,this thesis designs an intelligent grinding system for shoe soles based on monocular encoded structured light,and focuses on the 3D reconstruction technology and intelligent grinding path calculation in the system.Firstly,after considering the accuracy of the 3D reconstruction method and the measurement range of the hardware system,the monocular encoded structured light system based on phase shifting profilometry was selected as the 3D reconstruction system in this thesis.The internal and external parameters of the projector in the system are calibrated based on an improved three-wavelengths six-steps phase unwinding algorithm using horizontal and vertical encoding gratings.Compared with the traditional three-wavelengths six-steps method and three-wavelengths four-steps method,this method reduces the phase step error,decreases the nonlinear error,reduces the noise,and improves the phase decoding accuracy.The projector calibration using the method in this thesis is faster and more robust,and the reprojection error of the calibration results is smaller.Then,this thesis calculates intelligent grinding path points based on the upper boundary point cloud of the shoe sole.For the difficult problem of extracting the upper boundary point cloud from the complete point cloud of the shoe sole,this thesis proposed a point cloud upper boundary extraction algorithm based on three-coordinate scanning,which is used to calculate the entire point cloud of the upper boundary of the pre-grinding shoe sole.Based on the point cloud on the upper boundary of the sole,the intelligent grinding path points are calculated,and the intelligent grinding path is planned and smoothed.Finally,the robot equipped with grinding tools moves according to the grinding path converted by hand-eye coordinates and completes the intelligent grinding of the inner edges of the shoe sole.In order to verify the effectiveness and accuracy of the system in this thesis,experiments were conducted in industrial sites.The experimental results show that the accuracy of the projector calibration based on the improved three-wavelengths six-steps phase unwinding algorithm is 0.1715 pixels,the 3D reconstruction accuracy of the monocular encoded structured light system reaches 0.11 mm,the mean absolute error of the system’s hand-eye calibration result is around 0.3mm,which meets the precision requirement of ±0.5mm of the intelligent grinding system.Meanwhile,the intelligent grinding path calculation method in this paper is compatible with different types of soles,so that it can meet the flexible production requirements.For different soles of the same model,the algorithm generates an intelligent grinding path that matches exactly,which provides high robustness.