基于机器视觉的复杂背景下QR码检测

QR Code Detection Under Complex Background Based on Machine Vision

在未来制造业的发展中,智能制造将会是其重要的发展方向,而在工业上以QR码为基础的产品检测和分拣体系正在逐步完善。为了加强对流水线上产品信息的查验,如何快速高效地定位识别出产品上存储产品信息的QR码是一个亟待解决的问题。面对工业相机采集到工业流水线上产品的QR码图像存在如文字背景干扰、过度曝光、畸变、模糊等复杂背景的问题,提出一个改进的基于MQR的检测方法——OMQR。首先采用基于方向梯度直方图加支持向量机的方法对QR码分类,进而去除数据集中的负样本图像。然后对QR码图像做对比度增强、畸变复原等图像处理,最终完成对产品QR码的定位识别。基于MRP300机器视觉运动旋转实验开发平台,将本文方法和MQR码识别算法对比,结果表明:该检测方法能有效区分出工业相机拍摄的QR码图像,正确定位并识别出QR码信息,有效提高工业流水线下存在复杂背景的产品QR码检测准确性。

In the future development of manufacturing industry,intelligent manufacturing will be an important development direction,and the QR code based product detection and sorting system in industry is gradually improving.In order to strengthen the inspection of product information on the industrial assembly line,how to locate and identify the QR code of product information quickly and efficiently is an urgent problem to be solved.Faced with the complex background problems such as text background interference,overexposure,distortion,blur,and so on,an improved MQR-based detection method,called OMQR,is proposed in this paper.Firstly,the QR code is classified based on histogram of oriented gradient and support vector machine,and the negative sample image in the data set is removed.Then,the QR code image is processed by contrast enhancement,distortion restoration,and other image processing to implement the location recognition of the product QR code.Based on the MRP300,a development platform of machine vision motion rotation,the OMQR method is compared with the MQR-based method.The experimental results show that the proposed OMQR can effectively distinguish the QR code images captured by the industrial camera,correctly locate and recognize the QR code information,and effectively improve the QR code detection accuracy of products with complex background under the industrial assembly line.