一种快速、高效的车牌定位方法

Efficient and rapid approach for vehicle license plate location

基于车牌区域丰富的边缘特征,提出了一种改进的车牌定位方法。首先增强原始图像,并对原始图像和增强图像分别进行边缘提取;然后基于车牌区域边缘均匀、长短有限等特征,滤除背景及噪声边缘点;最后通过投影搜索出车牌区域。对采集到的各种复杂背景、环境下的车辆图像进行了大量实验,实验结果证明算法准确率高、适应性强。

License plate area contains rich edge and texture information,whieh is very uniform.This paper presents an efficient and robust approach for vehicle license plate location.Firstly,enhance the input image and detect the edges of the input image and enhanced image.Then remove most of the background and noise edges based on the threshold of the lengths of the edges. Finally,search the plate in the residual edge image through projection operation and segment it out from the original car image. The performance of the method was investigated on many real car images under complex conditions with various illuminations.The experimental results show that the proposed algorithms have high accuracy and robustness.