Research On Defect Detection Algorithm Of #0 Belt In Tire X-Ray Images

With the high-speed development of China’s transportation and auto industry,the consumption of tires increased rapidly.As to be the main component of a vehicle,tire plays a vital role in the safety of vehicles.In recent years,tire quality has become the key to get the core competitiveness of an enterprise,for the impacts of both improving of people’s vehicle safety ideas and the competitive market.There will be multiple defects in tire molding process.Tires with structural defects inside will lead to layer or even rupture as the carcass uneven force,which threats the safety of drivers and passengers seriously.The traditional human discrimination is subjective and intensive and there is lower rate of accuracy.Therefore,it’s urgent to put forward an effective algorithm for the automatic recognition of tire X-ray defects.The research in this paper is based on the inspection algorithm to the splice of #0 belt.The main creative achievements are as flows: According to the characteristics of the texture distribution of X-ray images and the gradation distribution of the defective areas,this paper proposed one inspection algorithm to the splice of #0 belt.This algorithm is divided into two parts:1.For non-periodic texture of tire X-ray images need to project,filter,calculate the relative error to detect the joint position,and then eliminate the joint area of upper and lower boundary effects,threshold segmentation,normalize in order to locate and quantify the joint defects in #0 belt.2.Searching for texture period blocks from a 45 degree projection of #0 belt in tire X-ray images.Then,determine the upper and lower boundaries of a block to find an appropriate standard block to match the splice block.Searching the boundaries helps locate and quantify the joint defects in #0 belt.This paper is the first time to solve the problem of the tire joint defects in #0 belt in recent years.The proposed algorithm is verified with GLCM and Gabor texture analysis algorithm on MATLAB platform,results show that this inspection algorithm can accurately locate the joint defects in #0 belt and quantify the size of the joint with 3.4% false positive rate and 2% false negative rate,which meets the industrial testing requirements.At last,this paper uses method of Visual C++6.0 and Visual Studio 2012 C# programming to design the tire X-ray automatic identification system and 5000 tire X-ray images are practical application tested,which is provided by MESNAC.The test results meet the requirements of defect detection in actual production.