车牌识别中倾斜牌照的快速矫正算法

在车辆牌照自动识别(LPR)过程中，车牌图像的倾斜会给下一步字符切分和识别带来困难。利用传统Hough变换进行矫正的运算量很大，文中提出采用轮廓跟踪方法去除牌照图像中和矫正无关的所有冗余信息并采用分级检测的策略，在最大程度上减少矫正过程的运算量。实验结果表明，该方法可以准确地检测出车牌的倾斜角度，同时矫正速度比传统Hough变换方法提高10倍左右。

A Fast Algorithm for License Plate Orientation Correction

In real-world License Plate Recognition (LPR), the plate tilt not only results in difficulties in character segmentation but also causes adverse impact to the final recognition rate. A direct use of the Hough transform is computationally very expensive. To overcome this difficulty, This paper presents two novel methods: one is to use the contour tracing algorithm to get rid of unrelated image regions to save computation; and the other is to use a hierarchical searching strategy in the Hough transform to gain further speedup. To demonstrate the effectiveness of the proposed algorithm, it conducts extensive experiments over a large number of real-world vehicle license plates. It reports a 10-time speed improvement.