高噪声条件下基于投影的二维条码倾斜检测

图像倾斜纠正是二维条码自动识别过程中的重要步骤,最常用的倾斜角检测方法是基于Hough变化的方法,实际应用中Hough变换峰值扩散效应显著,同时在速度上也难以适应实时应用的需求。提出使用投影法对倾斜的二维条码图像进行倾斜角判断,并通过加速策略提高检测速度,能够在高噪声条件下快速得到精确的结果。

A Skew Detection Method for 2D-Barcode Images in High Noise Condition Based on the Projection Method

The image skew correction is one of significant steps in the processing of 2D-barcode recognition. In general, the methods to detect images' skew angles are based on hough transforms, which cannot obtain accurate angle owing to the diffuse of its peak value, and because the hough transform is computationally expensive, it cannot apply to real-time applications. This paper proposes an approach that can obtain precise results in high noise based on the projection method, and also proposes several methods to speedup the detection process.