Research On Machine Vision System Of Detection Injection Molding Processing

Plastic products are widely used in various industries for its advantages.Injection molding processing is its main processing method.But many abnormal situations,such as product defects,product deformation,parts dislocation,are likely to occur in the process.It can lead to mold damage and waste of resources.Therefore,injection processing detection system based on machine vision was designed to detect the injection molding process.In chapter one,the background and significance of the project and development of machine vision detection technology were introduced,and the main contents of the paper was given.In chapter two,existing detection schemes were analyzed,and three detection functions of detection system were designed: initialization function,process monitoring and product inspection.The testing process of system and the relationship between the functions were introduced.In chapter three,the region segmentation system of the initialization functions,which could automatically extract tested product and recognize tested area,was set up.The principle of each pant of segmentation algorithms in the system was proposed and analyzed.Best value of parameters in segmentation system was got by experiments.Parameters adaptive mechanism was design to improve the stability of addressing environmental impacts.In chapter four,link recognition algorithm and residue detection algorithm of process monitoring functions and implementation flow process of monitoring were introduced.In chapter five,design and implementation process of product detection function were introduced.Implementation principle of objectives matching algorithm and target contour comparison algorithm were elaborated.Error compensation is designed to improve the accuracy of defects detection.In chapter six,the building process of system functions testing environment and contents of functional test experiment were introduced.The stability,the accuracy and timeliness of the detection system were verified and the detection ranges of the detection functions had got.The embedded platform was built,and the system was transplanted.In chapter seven,former chapters and pointed out the deficiencies were summarized.