基于边缘和水平集的复杂背景金属零件二维条码精确定位方法

针对由于背景复杂、光照不均等因素导致的复杂背景下金属零件上二维条码定位困难问题,提出一种基于边缘及水平集的复杂背景金属零件二维条码精确定位方法.首先,通过直线检测方法定位条码的L边;然后,采用一种改进的水平集图像分割方法对可能的条码区域进行图像分割,利用条码的边缘轮廓特征、灰度及梯度直方图双峰特征进行二维条码区域判断;最后,通过区域Hough变换精确定位二维条码区域.实验结果表明,所提出的方法平均定位准确率可达到95%,且其处理时间小于1s,适用于对大多数复杂背景下金属零件二维条码的定位.

Accurate Localization of 2D Barcode Marked in Mental Parts Under Complex Background Based on Edge and Level Set

Abstract： Aimed at the difficult location issue of two dimensional barcode marked in metal parts resulted from complex background and uneven illumination etc., this paper proposed an accurate localization method of 2D barcode marked in mental parts under complex background based on edge and level set. First, the “L” edges of 2D barcode was located by line detection. Next, an improved level set image segmentation was proposed to segment the probable barcode area, while the contour feature, gray and gradient histogram doublet features were used to judge if the probable barcode area had barcode or not. Finally, the 2D barcode area was accuratly located using Hough transform. The experimental result shows that the accuracy of the proposed method can reach up to 95%, and the processing time is less than 1 s, indicating that the proposed method is promising.