快速检测轴承表面缺陷方法的研究.pdfVIP

28 10 Vo l. 28 No . 10 2011 10 MICROEL ECT R ON ICS COM PU T ER October 2011 陶青平 吴锡生 ( 214122) : 轴承表面缺陷可分为两种 一种是低于目标灰度的缺陷 另一种是高于目标灰度的缺陷. 针对单 值分割 算法只能检测轴承前一种缺陷的局限性 而多 值分割算法仅能检测高于目标灰度缺陷的情况 提出了一种基于 OST U 的多次 值分割图像快速检测轴承表面缺陷的方法 该算法相对于多 值算法复杂度小 且根据轴承图像 的特点缩小了 值取值的范围 较好满足了在线检测轴承表面缺陷的实际需求. : 大津法; 多 值; 轴承; 缺陷检测; 灰度 : T P39 1. 4 : A : 1000- 7 180( 2011) 10 - 0098- 03 Rapid Detection Method of the Bearing Surface Defects T A O Q ing- p ing W U X-i sh en g ( Co llege of Co ntent N etw o rking Eng ineer ing J iangnan U niv ersit y W ux i 214122 China) Abstract: Bear ing surf ace def ect s can be divided into t w o classes one is the def ect of gr ay- scale lo w er th an the targ et and ano ther is the def ect of gr ay- scale hig er than the tar get . A im ing at the lim itatio n th at sing le- thr esho ld seg menta- tion alg or ithm can o nly detect the f ir st def ect and the situ atio n th at mult-i t hr esho ld seg m entation alg or ithm can only detect the second a new r apid detection m ethod o n the bear ing sur f ace def ect s is intro du ced b ased on multiple- OS- T U . T he complex it y of the algo rithm is sm aller th an mult-i threshold one and accor ding to the ch ar acter istics of the bear ing im age it nar ro w s the r ang e o f thresho ld valu es . So it satisf ies the need of o n- line detection of bear ing surf ace def ect s . Key words: O ST U ; mult-i thr esho ld seg m entat ion alg or ithm ; bearing ; def ect s detection ; gr ay- scale [ 2] 1