复杂背景下二维条码图像的提取

本文针对复杂背景下的PDF417条码，提出了一种合理的分割算法。首先使用彩色过滤算法对背景进行过滤，然后利用二次梯度算法提取图像中的正、负峰值，获取图像的分割阈值，使用这种方法后，可在最大程度上减少外界干扰。对于条码的倾斜，使用Hough变换得到图像倾角，再利用二维旋转矩阵旋转图像。为了减少目标区域的毛刺以及空白点，提出了一种良好的插值方法。确保旋转变换后的图像不失真。

The Extraction of 2D Bar Code Image Holding Complicated Background

This paper aiming at the PDF417 which has complicated background puts forward a logical segmentation algorithm to pick up the target district and uses colorful percolation method to filter the background first, Then use two-time grads algorithm to pick up the positive and negative peak-value and also obtain the specific value to dividing up the pictures. Using this method, we can re- duce the outside interference. For the gradient 2D bar codes, we use Hough transform to gain image obliquity and then employ 2D rotation matrix to revolve the pictures. In order to reduce the burr and the blank point of target district, a quite good algorithm is put forward to make sure that the pictures will not be fuzzy after rotation transform.