Fast Image Process and Detection of Distorted Blur-readable 2D Barcode

Blurring of two-dimensional (2D) barcode is very common in mobile Internet systems, which often leads to barcode recognition failure. In order to deal with the problem of barcode blur, blur-readable (BR) 2D barcode which can be decoded without deblurring is newly proposed. However, the research on the image processing of BR code are rare and the field of BR code detection and identification is almost blank so far, which impede the use of BR code in practical scenarios. In the application of BR code, distortion problems such as tilt and perspective of the image may lead to difficulty in identifying the BR code. In addition, in some industrial scenarios, we hope that the BR code can be instantly recognized to meet the real-time requirement. In this paper, an efficient image processing and detection method for BR code images with complex background is proposed. We use the ellipse feature of the finder patterns of distorted BR code to locate and correct the BR code images. The experimental results show that our BR code detection method is effective to detect distorted BR code images.