摘要
为了解决格子织物疵点检测的工程实际问题,提出一种基于周期分割模板减法的疵点检测方法。通过直方图匹配算法确定检测图像的纵向周期位置,以此为依据分割最小周期单元,后借助频域滤波及织物纹理周期提取完成图像滤波预处理操作,最后经过形态学操作处理和最大熵阈值分割算法作为疵点特征的提取依据。经过试验验证,该算法能够有效提取疵点的特征情况,并完成对疵点及其轮廓的识别检测与位置定位,具有较好的准确性与适应性。
In order to solve the practical engineering problem of defect detection of plaid fabric,a defect detection method based on periodic segmentation template subtraction was proposed.Through histogram matching algorithm,longitudinal period position of the detected image was determined.Based on this,the minimum period unit was divided.Then,the image filtering preprocessing operation was completed with the help of frequency domain filtering and fabric texture period extraction.Finally,morphological operation processing that performed and maximum entropy threshold segmentation were used as the basis for the extraction of defect features.After experimental verification,the algorithm could effectively extract the characteristics of defects and complete the identification,detection and position positioning of defects and their contours.It had better accuracy and adaptability.
作者
李辰一
于海燕
LI Chenyi;YU Haiyan(Donghua University,Shanghai,201620,China)
出处
《棉纺织技术》
CAS
北大核心
2023年第3期22-29,共8页
Cotton Textile Technology
基金
中央高校基本科研业务费专项资金资助。
关键词
周期分割
图像减法
格子织物
疵点检测
特征提取
periodic segmentation
template subtraction
plaid fabric
defect detection
feature extraction
