热敏打印机打印缺陷检测方法研究

Research on detection method of printing defect of thermal printer

针对热敏打印机生产过程需通过打印特定图案的标签以检测打印头是否存在缺陷,提出了一种热敏标签打印缺陷的视觉检测算法,用于检测由不合格打印机打印的缺陷标签,这些标签具有局部缺墨、坏道和区域不清晰等问题。首先,对采集到的标签灰度图像进行灰度值变换,以降低热敏纸颜色和光照的影响,并从图像中提取感兴趣区域(ROI);然后,针对常见缺陷,提取了各ROI的小波特征、投影特征和基于灰度共生矩阵(GLCM)的纹理特征。最后,利用支持向量机(SVM)对各ROI进行分类,并根据各ROI的分类结果得出最终的检测结果。研究结果表明:与人工检测相比,此算法的分类精度高、稳定性好,对缺陷样本有较高的召回率,基本满足生产过程中的检测需求。

Aiming at the problem that detecting the pattern defects caused by the defects of thermal printing head in production,a visual detection algorithm for thermal label printing defects is proposed,which can be used to detect defect labels printed by unqualified printers.These labels have some problems,such as local ink deficiency,bad channel and unclear area.First,The gray value of the collected label gray image is transformed to reduce the influence of background color and light,and the region of interest(ROI)is extracted from the image.Then,wavelet features,projection features and texture features based on gray co-occurrence matrix(GLCM)for each ROI are calculated.Finally,Support Vector Machine(SVM)is used to classify each ROI,and the final detection results are obtained through the classification results of each ROI.The experimental result shows that the algorithm has high classification accuracy and good stability,and has a higher recall rate for defect samples,which basically meets the detection requirements in the production process compared with manual detection.