一种改进的二维条码图像自适应阈值算法

摘要

For 2D barcode image with uneven collected illumination and obvious trough in its histogram,we present an image binarisation method,which uses wavelet multi-resolution analysis to carry out adaptive threshold.First,it obtains low frequency sub-image through wavelet decomposition,then it generates the dynamic threshold in local region by using a template of 2 ×3 pixels on low frequency sub-image.Finally it obtains global threshold according to the local threshold and thus achieves binarisation processing on PDF417 barcode image.Experimental results show that this method effectively reduces the impact of the uneven illumination on barcode identification,and is suitable for binarisation processing on 2D barcode images with complex background.%对于直方图有明显谷底和采集光照不均匀的二维条码图像，提出一种利用小波多分辨率分析进行自适应阈值的图像二值化方法。该方法首先通过小波分解获得低频子图，其次在低频子图上采用2×3模板产生局部区域的动态阈值，最后根据局部阈值得到全局阈值，进而实现PDF417条码图像的二值化处理。实验结果表明，该方法有效地减小了光照不匀对条码识别的影响，适合于背景较复杂的实时二维条码二值化处理。