基于IGD-IHT算法和PIQEDS-IBPSO-NESN算法的制动盘表面缺陷检测方法

打开文本图片集

SurfaceDefectDetection Method ofBrakeDiscsBased on the IGDIHT Algorithm and the PIQEDS-IBPSO-NESN Algorithm

Li Feng，Yu Zhen²³，Gao Juan4，An Qi5 （1.Haier School （Electromechanical Schol），Qingdao Technical Collge，Qingdao 266555; 2.Institute ofOceanographic

Instrumentation，QiluUniversityofTechnology（ShandongAcademyofSciences），Qingdao266O61;3.StateKeyLaboratoryof

PrecisioMeasuringTechologyandIstrument，TianjinUniversityTianjinO2;4.ScholofIformatics，Qingdoical College，Qingdao 266555;5.Departmentof Mechanical Engineering，Tsinghua University，Beijing100084）

Abstract：To improve therobustnessof traditional brakediscsurface defeet detection，anautomaticdetection instrumentbased onmachinevisionisdesigned.Thedefectfeatures of brakediscsare extractedusing the ImprovedGaussianDiffrencealgorithmandHough Transformalgorithm（IGD-IHT）.Anidentificationmethodfor brakedisc surface defectsisdesigned basedon the Perception-based Image QualityEvaluatorand DempsterruleimprovedBayesparticle swarm optimization-Nonlinear echo statenetwork to beteridentify defects.The experimental results show that accuracy of this method is more than 97% ，thefalse alarm rateisless than 1.5% and the missing alarm rate is less than 1.5% .The method described in this article is superior to traditional methodsand improves theacuracyof brakediscsurface defectrecognition.Factory testingshows thatthismethod canaccuratelyidentifyalmostall defects，and therearerelativelyfew falsepositivesor falsenegatives，and has high reliability and stability.

Keywords：Brake disc，Surface defect feature extraction，IGD-IHT algorithm，Surfacedefect identification，PIQEDS-IBPSO-NESNalgorithm

1前言

制动盘作为汽车制动系统的关键部件，通过摩擦保证制动效果[1-3]，一般由灰铸铁制成，铸造过程中，制动盘表面可能会出现沙眼、气孔等缺陷46]。（剩余16943字）