【作者】 朱新如； 金立左； 袁晓辉；

【Author】 Zhu Xinru;Jin Lizuo;Yuan Xiaohui;School of Automation,Southeast University;

【机构】 东南大学自动化学院；

【摘要】 全民扫码时代的到来日益增加了识别二维码的重要性,稳定、准确、快速检测二维码的算法是一项研究重点。QR(Quick Response)码是应用广泛的二维码,已有检测方法大多基于Hough直线检测或者基于位置探测图形检测,检测率低且难以检测多码。本文提出了在复杂背景下使用级联分类器检测QR码的算法,不仅可以对一张图像进行多码检测,还具有更高的检测率。在对图像进行预处理后,以Viola-Jones快速目标检测框架为基础,利用Harr-like特征训练级联分类器检测包含QR码的候选区域,再根据QR码位置探测图形的结构对其进行准确定位。实验结果证明本文算法的可行性与有效性。更多还原

【Abstract】 The arrival of the era of universal scanning code has increasingly increased the importance of recognizing twodimensional codes. The algorithm for stable, accurate and rapid detection of two-dimensional codes is a research hotspot. QR(Quick Response) codes are widely used two-dimensional codes. Existing detection methods are based on Hough line detection or finder patterns detection. The detection rate is low and it is difficult to detect multiple codes. This paper proposes an algorithm for automatic extraction of QR codes using a cascaded classifier in complex background. It can not only detect multiple QR codes in one image, but also has a higher detection rate. After the image is preprocessed, based on Viola-Jones’ s rapid target detection framework, the candidate region with QR code is detected by training cascade classifier using the Haar-like features, it is accurately positioned according to the internal structure of the QR code. Experimental results prove the feasibility and effectiveness of the proposed algorithm.更多还原