节点文献

遗传算法在车牌识别系统中的应用研究

Research on the Application of Genetic Algorithm in License Plate Recognition System

【作者】 徐壮飞

【导师】 曹杰;

【作者基本信息】 南京航空航天大学 , 信号与信息处理, 2009, 硕士

【摘要】 近年来我国城市智能交通系统发展得很快,车牌识别系统作为城市智能交通系统中信息采集的一种手段,也得到了很快的发展。遗传算法是一种模拟自然界进化过程的寻优算法。自提出之日起,遗传算法的理论得到了很大程度的完善,出现了许多针对不同问题的改进型遗传算法。由于遗传算法有着解集搜索能力、寻优能力、容错能力、适应能力、隐含并行性等特点,在很多学科领域尤其是图像处理领域内,都有着很大的研究价值和实用价值。本文在详细研究了遗传算法原理和国内外车牌识别系统的基础上,结合我国车牌特点提出了一套新型车牌识别系统,取得了较好的效果。本文对车牌识别技术的研究,主要的工作有三个方面:1深入研究了遗传算法的原理及改进思想,并根据车牌区域特征适当构造适应度函数,成功地将遗传算法应用到车牌定位之中,实验效果良好;2针对车牌定位、倾斜校正和字符分割部分的各种主流思想的不足,提出了较为行之有效的改进算法;3深入学习了人工神经网络的基本原理,并在此基础上结合模板匹配思想实现了车牌字符的分层识别算法。

【Abstract】 Intelligent Transportation Systems (ITS) is progressing rapidly in recent years in China. At the same time, as the information collection method for ITS, the license plate recognition technoloty has also developed very fast.The genetic algorithm (GA) is a kind of optimization algorithm which imitates the process of the natural evolution. From the day on which GA was proposed, the theory of GA has obtained a refinement to a great extent, and a lot of improved genetic algorithms have been put forward aiming at different questions. Because of the abilities of GA such as solution set searching, optimization, fault tolerance, adaptation, potential parallelisom, there are great research value and utility value of GA in many fields especially in image processing.Based on the detail analysis for the genetic algorithm (GA) and the vehicle plate recognition systems (VPRS), a new VPRS put forward in this paper has achieved good effects, concerning the characteristics of Chinese vehicle plate.The research presented here for VPRS is characteristic of the three creation points showing bellow:First, by researching into GA, a new algorithm has achieved satisfactory effects in vehicle plate location, in which the GA has been applied successfully.Second, aiming at the shortages of mainstreams in vehicle plate location, tilt correction, and character segmentation, some workable improved algorithms are proposed.Last, based on the intensive study of artificial neural network (ANN), a stratified algorithm of character recognition is realized, combining with the thorey of template matching.

节点文献中: 

本文链接的文献网络图示:

本文的引文网络