智能停车场管理系统的研究与设计

发布时间：2018-05-26 23:25

本文选题：多车牌定位 + 车牌识别　；参考：《青岛科技大学》2017年硕士论文

【摘要】：目前机动车的数量正在与日俱增,与之共同增长的还有机动车的停车需求,这对停车场管理提出了更高的要求。而现在的停车场管理系统存在重视收费功能、缺乏有效的监控措施等问题,随着移动互联网的发展,司机也需要一款可以在手机上就可以查找停车位的软件。因此,本文设计了适用于大型停车场的智能管理系统,其中多车牌识别算法在智能停车场管理系统中具有至关重要的作用,智能停车场管理系统中的诸多功能依赖于多车牌识别算法的识别结果。本文主要研究内容如下:(1)详细阐述了车牌识别的各个步骤中常用的各种算法,并对比了算法之间的优点与缺陷,然后吸取各种算法的精髓,提出了一种基于多算法融合的多车牌定位方法。首先,为了不遗漏潜在的车牌区域,分别使用基于最大极值稳定区域的文字定位、颜色检测与边缘检测方法粗选车牌候选区域;其次对车牌候选区域做倾斜矫正处理;最后提取车牌候选区域作特征信息,采用支持向量机来精确定位车牌区域。实验表明,本文提出的多车牌定位方法准确率高,抗干扰能力强。(2)基于多车牌定位方法的定位结果,本文使用基于连通域的方法对车牌字符进行分割操作。最后,本文使用BP神经网络算法对车牌字符进行识别,从而实现了对多车牌的识别,该算法使用C++语言与OpenCV实现。实验表明,本文的多车牌识别方法准确率高,具有识别多种颜色车牌的能力。(3)本文提出了适用于大型停车场的智能管理系统设计方案,系统服务器使用Java语言实现,服务器端使用socket与多车牌识别算法通信,手机端使用OC语言实现。系统设计了车位预约、出入口控制、停车计费以及车辆查找等模块。(4)本文实现具有车位预约、停车付费与车辆查找等功能的手机软件。用户通过手机软件查看周边的停车场的空余车位信息,并对其满意的停车场进行预定;基于车牌识别系统,系统可以精确的计算停车时间与费用,用户手机与车辆绑定后可以在手机完成支付停车费用。本系统可以在手机上使用,通过智能手机预约车位、结算停车费,更符合现代生活方式,同时提高了停车场的运行效率。
[Abstract]:At present, the number of motor vehicles is increasing day by day, with which the parking demand of motor vehicles is increasing, which puts forward higher requirements for parking lot management. With the development of mobile Internet, drivers also need a software that can find parking space on mobile phone. Therefore, this paper designs an intelligent management system for large parking lots, in which the multi-license plate recognition algorithm plays an important role in the intelligent parking lot management system. Many functions of intelligent parking lot management system depend on the recognition result of multi-license plate recognition algorithm. The main research contents of this paper are as follows: (1) the common algorithms in every step of license plate recognition are described in detail, and the advantages and disadvantages between the algorithms are compared, then the essence of the various algorithms is absorbed. A multi-license plate location method based on multi-algorithm fusion is proposed. Firstly, in order not to omit the potential license plate region, the text location based on the maximum extremum stable region is used respectively, the color detection and edge detection method select the license plate candidate region extensively; secondly, the tilt correction of the license plate candidate region is done. Finally, the license plate candidate region is extracted for feature information, and support vector machine is used to accurately locate the license plate area. Experimental results show that the proposed multi-license plate location method has high accuracy and strong anti-jamming ability. (2) based on the location results of the multi-license plate location method, the method based on connected domain is used to segment the license plate characters in this paper. Finally, this paper uses BP neural network algorithm to recognize the characters of license plate, and realizes the recognition of multiple license plates. The algorithm is implemented in C language and OpenCV. Experiments show that the method of multi-license plate recognition in this paper has high accuracy and has the ability to recognize multi-color license plates. In this paper, an intelligent management system for large parking lot is proposed. The system server is implemented in Java language. The server uses socket to communicate with the multiple license plate recognition algorithm, and the mobile phone uses OC language to realize. The system designed parking reservation, entrance control, parking billing and vehicle search module. 4) this paper realizes the mobile phone software with parking reservation, parking payment and vehicle search functions. Through mobile phone software, users can check the free parking space information in the surrounding parking lot and book a satisfactory parking lot. Based on the license plate recognition system, the system can accurately calculate the parking time and cost. The user's mobile phone can be connected to the vehicle to pay the parking fee. The system can be used on the mobile phone, through the smartphone booking parking space, the settlement of parking fees, more in line with the modern way of life, while improving the efficiency of the parking lot.
【学位授予单位】：青岛科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP391.41;TP311.52

【参考文献】