Research On Online Stereo Information Recognition System Of Shoe Mid-sole Based On Machine Vision

Most of the midsole is die-cast in a mold using a soft,ultra-light,and elastic foam.During the mid-floor production inspection process,workers need to sort and follow the information on the midsole.The midsole is die-cast with important information such as shoe size,model number,left and right feet,and the information is die-cast in the three-dimensional character or the stereoscopic two-dimensional code on the surface of the midsole.At present,the method of manually identifying information on the midsole of the shoe and measuring the size is cumbersome and time consuming,and the method of measuring the size using the 2D image processing is not accurate enough.In order to realize the automatic on-line detection based on the midsole of shoes,this paper develops the online stereo information recognition(stereo character recognition,stereo two-dimensional code recognition and size measurement)system based on machine vision technology,and studies the key points.Analysis,the main research results of the paper are summarized as follows:(1)According to the actual production environment of the midsole of the shoe,combined with the requirements of the key test indicators proposed by the buyer,the technical implementation plan of the shoe midsole detection system is designed.(2)According to the hardware and software system scheme of the shoe midsole detection system and the shape and structure characteristics of the midsole of the shoe,the imaging experiment platform is built,and the system hardware scheme is designed to be reasonable.(3)According to the hardware design scheme of the detection system,the algorithm research of stereoscopic two-dimensional code,stereo character recognition and shoe midsole size measurement is carried out.(4)Based on the completion of the shoe midsole two-dimensional code recognition and the stereo character recognition algorithm,the midsole software design of the midsole recognition system is carried out.(5)Test the hardware and software performance of the shoe midsole stereo information recognition system,analyze the results according to the results,andimprove the hardware and software system.In the final system,the speed can be4s/only,the two-dimensional code,character and length recognition have better recognition effect,the system runs stably and the recognition rate reaches 95%.(6)Propose the existing problems and future research directions,so as to improve the whole system in the next step and improve the adaptability of the system in the actual working environment.