基于滤波-还原的二维条码识别投影算法

给出了一种基于PDF417二维条码结构特性的算法,该方法设计滤波器滤除噪声部分,再采用点搜索法进一步去除不符合条码边界特点的孤立的点和容易与边界混淆的竖直划痕。对条码图像进行复原,由于基本清除了污染,污染部分找不到复原的种子,不会被恢复。该方法可以去除滤波对有效条码边缘的影响,同时保证图像的可靠恢复。对于恢复后的条码图像采用改进的边缘检测投影方法译码。实验结果表明该方法具有良好的性能,可以降低对条码纠错等级的要求,提高了译码的效率和准确率。

Two Dimensional Bar Codes Decoding Algorithm Based on Projection and Filtering-Restoring Principles

The paper presents an algorithm based on PDF417 two dimensional bar codes' structure. It first proposes a filter to eliminate noise, then deals with vertical scratches and those points that do not comply with the bar codes' boundary characteristics by a point-search strategy. Later, during process of restoring the bar coded pictures, the contaminations will not be restored because-their seeds for restoration have been cleaned beforehand. This algorithm features a reliable picture restoration that is free from the negative consequences imposed on the effective bar code edges by filtering. The restored pictures can be decoded by the improved edge-detecting projection method. Experiments have proved that this algorithm improves efficiency and accuracy of the decoding process while demanding a lower error correcting level.