基于边沿像素定位的金属表面破损视觉检测算法  

Metal Surface Damage Visual Detection Algorithm Based on the Edge Pixel Positioning

作　　者：樊雄[1,2] 

机构地区：[1]广西大学材料科学与工程学院,南宁530001 [2]广西工业职业技术学院,南宁530001

出　　处：《科技通报》2013年第4期196-198,共3页Bulletin of Science and Technology

摘　　要：利用像素分割算法进行金属表面破损视觉检测,在图像采集的过程中会受到采集空间光照条件的影响,将造成金属表面像素错位的缺陷,从而导致检测准确性降低。为此,提出了一种基于边沿像素定位算法的金属表面破损视觉检测方法。采集金属表面图像,利用单极化方法对边沿像素进行提取,通过边沿像素定位方法计算边沿像素的空间位置,从而实现金属表面破损视觉检测。实验证明,这种算法能够有效提高金属表面破损视觉检测的准确性,取得了令人满意的结果。Use pixel segmentation algorithm for metal surface damage visual inspection, in image acquisition process will be affected by acquisition space illumination condition the influence, wilt cause the metal surface pixel displacement of defects, resulting in a decrease in accuracy detection. Therefore, this paper proposes a based on the edge pixel localization algorithm of metal surface damage visual inspection method. Acquisition metal surface image, using the method of edge pixels ChanJiHua were extracted and through the edge pixel location method to calculate the edge pixel space position, so as to realize the metal surface damage visual inspection. Experiments show that this algorithm can effectively improve the metal surface damage visual inspection accuracy, and achieved satisfactory results.

关 键 词：边沿像素定位 金属制造业 视觉检测 

分 类 号：TG115[金属学及工艺—物理冶金]