【作者】 黄战华； 刘正； 张振； 蔡怀宇； 杨基春； 张尹馨；

【Author】 Huang Zhanhua Liu Zheng Zhang Zhen Cai Huaiyu Yang Jichun Zhang Yinxin (College of Precision Instrument and Opto-electronics Engineering, Tianjin University; Key Laboratory of Opto-electronics Information Technology Science, EMC, Tianjin 300072, China)

【机构】 光电信息技术科学教育部重点实验室 天津大学精密仪器与光电子工程学院；

【摘要】 针对工业现场实时性要求,研究了一种轮胎X光纹理缺陷的快速检测方法。根据X光成像原理,轮胎X光图像中的纹理基元中心处灰度值最小,通过差分方法查找局部区域极值点并进行搜索识别提取纹理骨架,然后利用游程编码方法对纹理骨架进行矢量化。在纹理矢量图形的基础上建立纹理的特征描述模型,通过特征描述模型检测和识别纹理缺陷。实验结果表明,该方法能快速地检测和识别出轮胎纹理中的各种结构缺陷,检测速度比在位图中提高了70%以上。更多还原

【Abstract】 To meet the command of real-time performance in industrial field,a method has been researched for fast detecting texture defects in X-ray tyre image.Based on the principle of X-ray imaging,the gray level at the center of texton is the least in local area.A difference method was adopted to find local extreme points.The skeleton of texture was acquired through searching and recognizing these points.The run-length coding method was adopted to vectorize the texture skeleton.In texture vector graph,the character description model was founded,and through this model,texture defects were detected and recognized.The experiment results showed that structure defects in tyre texture were fast detected and recognized,and the detection time was reduced 70%.更多还原