车牌自动识别系统应用研究

The Overview of Vehicle License Plate Recognition

通常,人们能够很容易通过人眼读取字符而计算机却不能,除非计算机有智能处理能力。VLPR(车牌识别系统)就是用来处理车牌定位和车辆号码识别的智能系统。VLPR主要有三大功能模块:车牌提取、分割和识别。车牌提取功能主要通过边缘检测实现。车牌字符分割部分主要使用过滤、稀释及水平垂直投影技术实现。车牌识别技术是VLPR系统最后一步,主要是识别车牌中的字符和数字。车牌识别系统的实现取决于输入图像的质量,一些外部环境造成的影响因素,如阴影、雨痕和模糊也会降低VLPR系统性能。为此,主要分析多种VLPR方法,对基于特征提取和分类表的VLPR文献进行分类总结,并对车牌字符的检测、定位和识别进行阐述。

In general, human can easily read the character but the computer cannot read the character easily unless the computer is pre-trained to do so. VLPR deals with License Plate localization and vehicle number recognition. VLPR has three modules namely License Plate Extraction, Segmentation and Recognition. Extraction of plate is mostly carried out using edge detection. In Segmentation various tasks such as Filtering, Thinning, Vertical and Horizontal projections are performed. Recognition is the final stage to recognize character and numbers. The fulfillment of the Automated License Plate recognition system depends on quality of the input image. The performance of VLPR is degraded due to noise such as shadow, occlusion, rain streaks and blur due to vehicle speed. This paper analyzes various VLPR methods, classifies the VLPR literature based on feature extraction and classification schemes, presents the survey of approaches for detection, localization and recognition of plate character ＆ numbers.