色度亮度空间中的线束端子显微图像分割

发布时间：2019-06-18 16:01

【摘要】：为提取检测线束端子质量的相关数据,提出一种在色度亮度空间中分割线束端子显微图像的算法。该方法利用线束端子显微图像的颜色分布特点,用阈值法分割线束端子。通过分析线束端子显微图像的颜色特点,把彩色图像转换至色度亮度空间。为了保证图像内部的颜色均匀,对其中的色度向量进行滤波。借助获取色度向量的红色通道,考虑红色通道直方图双峰夹一谷和双峰夹多谷情况下阈值并分割图像,使用数学形态学方法处理二值图像获得内径轮廓。实验结果表明,分割得到的端子内径轮廓周长偏差小于0.1%,表明此方法可以准确地分割线束端子显微图像。
[Abstract]:In order to extract the related data to detect the quality of wire harness terminal, an algorithm for dividing the microscopic image of wire harness terminal in chromaticity luminance space is proposed. Based on the color distribution characteristics of the microscopic image of the wire harness terminal, the threshold method is used to segment the wire harness terminal. By analyzing the color characteristics of the microscopic image of the wire harness terminal, the color image is converted to the chromaticity brightness space. In order to ensure the uniform color of the image, the chromaticity vector is filtered. With the help of obtaining the red channel of chromaticity vector, the threshold of the red channel histogram is considered under the condition of double peak clip one valley and double peak clip multiple valleys, and the image is segmented. The mathematical morphology method is used to process the binary image to obtain the inner diameter profile. The experimental results show that the circumferential deviation of the inner diameter profile of the segmented terminal is less than 0.1%, which shows that this method can accurately segment the microscopic image of the wire harness terminal.
【作者单位】：河南理工大学计算机科学与技术学院;广东省数据科学工程技术研究中心;
【基金】：国家自然科学基金(U1404103) 河南省教育厅科技攻关重点项目(14A520029)
【分类号】：TP391.41

【相似文献】