几何特征形态学车牌识别系统研究  

Research on Geometry Morphology License Plate Recognition System Based on LabVIEW

作　　者：李雄[1] 裴承鸣[1] 郑华[1] 

机构地区：[1]西北工业大学数据处理中心,陕西西安710072

出　　处：《计算机仿真》2012年第4期353-356,370,共5页Computer Simulation

摘　　要：优化识别车牌识别问题,由于图像中的环境背景受到天气、照明等因素的影响,车牌定位不清。为解决上述问题,提出了一种结合车牌图像自身几何特征的数学形态学车牌识别系统的方法。首先应用直方图的灰度增强和局部阈值算法对车牌图像进行的预处理,通过对比采用梯度算子Roberts对图像进行有效的边沿检测,再根据提出的几何特征形态学车牌定位识别方法对灰度车牌图像进行车牌区域精确定位,采用模板匹配和神经网络方法实现字符识别。通过对实际场景中车牌图像样本进行仿真,证明了上述方法的有效性,且借助于实时性好的LabVIEW平台,较好地实现车牌识别定位优化问题,为实际交通管理提供了依据。This paper proposed a method of Geometry morphology license plate recognition system based on the LabVIEW.First,the gray-scale enhancement and local threshold algorithm were applicated for the license plate image pre-processing based on histogram.By contrast,the Roberts gradient operator was used for image edge detection.According to the geometric characteristics of morphological License Plate Recognition method presented in the paper,the precise plate area of gray-scale images of license plates was located,and the template matching and neural network were used for character recognition.The simulation with actual scene vehicle images samples proves the effectiveness of this method.And by means of LabVIEW,the algorithm has a better application value in Engineering.

关 键 词：图像处理 车牌定位 数学形态学 字符识别 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]