基于Halcon的车牌识别技术研究
基于 Halcon 的车牌识别技术研究
摘要
汽车牌照识别系统是智能交通系统的核心部分,可用于公路电子收费出入控制和车流监控等众多场合,它主要包括车牌定位、字符分割、文字训练和识别四个部分。其中,车牌的定位、分割和训练对后续的字符识别有重要影响。本文在总结近年来国内外在车牌定位和分割领域的最新研究进展的基础上,对车牌定位及分割的算法进行了系统的研究。
在车牌定位上,选择 RGB 三通道进行处理,将蓝的车牌区域
初步分割出来。通过面积、形状、阈值等参数的限定,进一步处理。
对区域进行选择和排除,最后得到完整的车牌图像。在文字训练过程中,通过大量的程序调试,最终成功的识别并显示了车牌号码。
最后,通过 Halcon 软件生成 VC 程序,构建了车牌定位及字符分割系统的软件平台。大大节省了编程以及调试带来的麻烦。对实际
获取的车牌进行了程序调试,获得了令人满意的结果,为其它车牌的
识别创造了条件。
关键词:车牌定位,字符分割,文字训练,OCR
III
基于 Halcon 的车牌识别技术研究
Abstract
Vehicle license plate recognition system LPRS is the hardcore of the
intelligent traffic system ITS .It can be used in the fields such as electron
charging,pass controlling,automobile stream supervising and so on.LPRS
consists of four main modules which are license plate location ,character
segmentation,character trains and character recognition.The former three
modules have great effects on the last module. This article in summary
in
recent    years    domestic    and    foreign  in  car  license  localization    and  in
division domain newest research development foundation, has
汽车牌照安装
conducted
system's research to the car license    localization    and    the  division
algorithm.
In the stage of license plate location, the blue license plates separate
regional preliminary by RGB three-channel. Further processing is limited
by    the  parameters      of  the  area,  shape,    threshold    and  so  on.  After    a
selection    of  regional  and  exclusion,we      get  the complete  license  plate
images. Through a lot of debugging process, the ultimate success of the
identification    and    show    the  license  plate    number
in  the  process    of
training.
Finally, the car license localization and
the  character    division
system's  software  platform  has  been  constructed  by
the  VC  procedure
through the Halcon software. The result shows the algorithm in this
paper
is effective after the actual access to the plate for the program debugging
and prepares good conditions for the continued license plate recognition.
Keywords:license plate location,character        segmentation,Character