CAJ | 학술논문

针对织物破洞、断经、跳花和浆斑这4种类型疵点的特征，提出一种基于图像信息的检测方法。根据织物疵点的几何特性选择结构元素，应用形态学运算突出疵点形状信息，应用灰度积分投影法确定织物疵点的位置信息，然后分析纹理特征，提取织物疵点面积、周长等特征参数。结果表明：本方法克服了经典边缘检测算子只能给出疵点轮廓的弊端，利用图像信息检测进行计算机辅助验布，对于纺织信息化具有一定的理论意义和实用价值。
침대직물파동、단경、도화화장반저4충류형자점적특정，제출일충기우도상신식적검측방법。근거직물자점적궤하특성선택결구원소，응용형태학운산돌출자점형상신식，응용회도적분투영법학정직물자점적위치신식，연후분석문리특정，제취직물자점면적、주장등특정삼수。결과표명：본방법극복료경전변연검측산자지능급출자점륜곽적폐단，이용도상신식검측진행계산궤보조험포，대우방직신식화구유일정적이론의의화실용개치。
A detecting method based on image information is proposed for four kinds of fabric defects (holes, cracked ends, skips and hard size) detection. The structure elements are selected according to the geometric characteristics of fabric defects, the morphological operation is used to highlight the shape information of fabric defect and grayscale integral projection method is used to determine the location information of fabric defect. Then texture features are analyzed to extract characteristic parameters such as area of fabric defects and perimeter. The result shows that the proposed method overcomes the problem that the classical edge detection operators can only give profile of fabric. Meanwhile, the proposed method uses image information detection for computer-aided cloth in-spection, which has theoretical significance and practical value for the textile information.