The Software Design Of The Wheel Automatic Detection Based On X-ray

Wheel,is the vital component of safety driving.Each country is very strict with the wheelquality,so the wheel must be detected before they leave the factory.Because of the X-raynondestructive testing has the advantages of convenient direct and quantitativedetermination,so it’s an effective tool of the modern industry development. While thetraditional detection method is rated a personnel by means of a fluorescent lamp,scale and amagnifying glass to be done manually.This method is not only affected by human factors,butaslo the production efficiency is low.So,it is difficult to meet the need of the on-linedetection.The x-ray detection methodology is pushed forward by the improvement ofProduction efficiency and quality requirements in modern industry.In this context,in order todetect the defects accurately and overcome the shortcomings of artificial detection.Theautomatic recognition of defects detection technology based on image understanding hasbecome an important development direction in the field of X-ray detection.This topic doesn’t only research and analyze the technology of the wheel automaticalidentify,but also design the process of X-ray automatic detection and identification system.Atlast,it developed a set of automatic recognition defect detection software based on X-ray realtime imaging of automotive aluminum wheel. The software consists of five modules: awheel-type identification and process file test school module,the image acquisitionmodule,the target detection module,the data exchange module and the output module.In thefirst module,first it takes pictures of a new model wheels automatically using the PanasonicPV500 machine vision system and stored in the system as the image of the wheel modeltemplate,then making process of the wheel of this new model file.Next,identify the whleel-type about the line wheels.And then call the stored process file of the wheels-typing todetect the wheel.It uses the Intel open source image processing library OpenCV to write theimage algorithm in the target detection.First it proposed an improved de-noising algorithm ofthe recursive filtering for real-time imaging which Combines the time domain and the spatialfiltering.Experiments show that the algorithm can reduce the noise in real-time imaging.Andthen it presents the selection and preservation of an ROI region on human-computerinteraction which is used in subseque image defects recognition processing.It isn’t only reducethe interference of the unrelated background area and facilitate the detection of the defectregion,but aslo reduce the processing time.Finally,it uses the method based on edge andregion information to extract the wheel defect.Through the engineering verification,this software is running well and it can completethe process of wheel detection fully automatic.But the algorithm aslo need to be improved,sothat it can reduce the rate of false positives and missing of all types of defects as much aspossible to make it become more perfect and to meet the requirements of industrialapplications.