Applying QR code to trace the source of products is conducive to improving the safety of their quality. Because of the influence of uneven illumination, it is difficult to recognize QR code fast and exactly. To solve this problem, a fast and efficient method for processing QR code image of uneven illumination was proposed. Three exper-imental researches about QR code image of uneven illumination were carried out. Firstly, Otsu algorithm’s binary effect for original images and their four blocks, nine blocks, sixteen blocks was studied; secondly, Otsu algorithm’s binary effect after Retinex image enhancement for original images was studied; finally, Otsu algorithm’s binary effect after histogram equalization for original images and their four blocks, nine blocks, sixteen blocks was studied. The experi-mental results show that when uneven illumination QR code image is divided into sixteen blocks, the average recogni-tion rate of histogram equalization method is 94.4%, which is 43.3% and 10% higher than Retinex enhancement method and Otsu method respectively. Moreover, the average operation time of histogram equalization method is 0.14 seconds, which is 0.004 seconds more than Otsu algorithm and 0.139 seconds fewer than Retinex method. Thus, it could be con-cluded that QR code images under the influence of uneven illumination can be identified fast and effectively by the his-togram equalization method..%QR二维码应用于商品溯源有助于提升商品的质量安全水平，但因其易受光照不均的影响，导致QR二维码难以快速准确识别，为了解决这一问题，本研究提出了一种光照不均 QR 二维码图像的快速高效处理方法。以光照不均QR二维码图像为研究对象，首先研究Otsu算法对其及其4分块、9分块、16分块的二值化效果；其次研究Retinex图像增强算法对其增强后Otsu算法的二值化效果；最后研究QR二维码图像及其4分块、9分块、16分块在直方图均衡化后Otsu算法的处理效果。实验表明，当把光照不均QR二维码图像分为16块后采用直方图均衡化方法处理时，平均识别率为94.4%，与Retinex图像增强方法和Otsu算法相比，分别提高了43.3%和10%，平均运行时间为0.14秒，比Otsu 算法略慢0.004秒，比 Retinex 方法快0.139秒。表明采用直方图均衡化方法能够有效地实现光照不均影响下 QR二维码图像的处理。
展开▼
