| 高清整皮表面缺陷自动检测与定位 |
| |
| 引用本文： | 范大煌,丁 磊,邓杰航.高清整皮表面缺陷自动检测与定位[J].教育技术导刊,2019,18(12):146-150. |
| |
| 作者姓名： | 范大煌 丁 磊 邓杰航 |
| |
| 作者单位： | 广东工业大学 计算机学院，广东 广州 510006 |
| |
| 基金项目： | 国家自然科学基金项目（61202267） |
| |
| 摘 要： | 针对传统CCD皮革检测系统成像视场小，以及传统皮革缺陷检测方法主要对皮革局部区域进行检测，应用在整张皮革缺陷检测上还需进行图像拼接等问题，提出一种高清整张皮革表面缺陷自动检测方法。首先根据高清整皮一次成像系统获取的整张皮革（整皮）图像特点，提出基于饱和度的整张皮革有效区域提取方法提取皮革区域（Region of Interest， ROI）；然后，为了在整皮图像ROI区域内一次对多种不同类型缺陷进行检测与定位，提出基于增强缺陷边缘的算法对皮革表面缺陷进行自动检测与定位。与传统皮革局部区域缺陷检测方法相比，该方法直接对整皮图像进行检测，无须再应用图像拼接算法得到最终结果，且能在一定程度上克服光照不均的影响，因此可以快速、准确地自动检测与定位整张皮革缺陷。
|
| 关 键 词： | 整皮 皮革缺陷 数学形态学 多尺度积 缺陷检测 |
| 收稿时间： | 2019-02-26 |
Automatic Detection and Localization of Surface Defects for a Whole Piece of Leather Image of High Definition |
| |
| FAN Da-huang,DING Lei,DENG Jie-hang.Automatic Detection and Localization of Surface Defects for a Whole Piece of Leather Image of High Definition[J].Introduction of Educational Technology,2019,18(12):146-150. |
| |
| Authors: | FAN Da-huang DING Lei DENG Jie-hang |
| |
| Institution: | School of Computer Science， Guangdong University of Technology， Guangzhou 510006，China |
| |
| Abstract: | In the view that the imaging field of view is small on the traditional CCD leather detection system， traditional leather defect detection methods are mainly used for defect detection in local areas of leather. However， this method is applied to the detection of defects in the entire leather. This also has problems such as image stitching. An automatic detection method for high-definition whole leather surface defects is proposed. Firstly， in according with the characteristics of the entire leather image obtained from the high-definition one-shot imaging system， an efficient region extraction method based on saturation is proposed to extract the region of Interest （ROI）. Then， in order to detect and locate many different types of defects in the ROI area of the whole skin image， an algorithm based on enhanced defect edge is proposed. It can automatically detect and locate the surface defects of leather. To compared with the traditional method， the method can directly detect the whole leather image. It diesn’t need to apply the image stitching algorithm. Moreover， the method can overcome the influence of uneven illumination to a certain extent. On the experiment， the method proposed in this paper can quickly and accurately detect and locate the defects of the whole leather. |
| |
| Keywords: | whole piece of leather leather defects mathematical morphology multi-scale product defect detection |
|
| 点击此处可从《教育技术导刊》浏览原始摘要信息 |
| 点击此处可从《教育技术导刊》下载免费的PDF全文 |
