射频识别与视频识别在智能交通中的应用研究

发布时间：2018-05-20 11:10

本文选题：车辆识别 + 射频识别　；参考：《兰州理工大学》2014年硕士论文

【摘要】：随着物联网技术的兴起,以及智能交通系统的高速发展,自由行驶车辆的精确定位成为智能交通科研领域的一大研究热点,截止当前,车辆的精确定位技术在不停车收费系统、智能停车场系统、智能小区门禁系统等多个领域都展现出了较为显著的应用效果。伴随着智能交通系统的快速发展,以往基于射频识别或者视频识别的单一模式的识别技术已经无法满足日益复杂的道路状况对车辆识别系统的精确性和鲁棒性的要求。为此,本文主要就射频识别和视频识别中的标签防碰撞问题以及车牌定位问题展开了深入的研究和探讨,主要研究工作如下： 1、针对射频系统中现有的电子标签二进制搜索类算法吞吐率较低的弱点,本文对现有的二进制搜索算法的编码方式进行了改进,得到了一种改进的二进制搜索优化编码算法,该算法利用标签序列号的唯一性和曼彻斯特编码可以准确的识别标签碰撞位置的特性,将阅读器发送的字符进行二进制优化编码,改进后的算法显著的降低了搜索的次数以及发送的字符的长度,使得射频识别系统的效率得到了提升。 2、针对光照的角度和距离不同,车辆运动速率不一,多变、复杂的背景等客观因素在对比度方面对图像质量构成的干扰,本文对现有的小波变换算法进行了改进,得到了一种改进的基于小波变换的车牌定位算法,该算法对所拍摄的车牌图像在角度变化、距离变化以及光照强度变化的情况下,定位效果较好,特别适宜于自然场景识别。理论分析和仿真结果均表明,本文提出的改进算法是可行的。二进制搜索优化编码算法在系统吞吐率、传输数据长度以及搜索次数三个方面的性能明显优于传统的二进制搜索和动态二进制搜索算法；基于小波变换的车牌定位算法在图像拍摄角度大于30度、拍摄距离大于3米的情况下均取得了较好的定为效果。
[Abstract]:With the rise of the Internet of things technology and the rapid development of intelligent transportation system, the precise positioning of free moving vehicles has become a research hotspot in the field of intelligent transportation research. Up to now, the accurate positioning technology of vehicles is in the toll collection system without stopping. Intelligent parking system, intelligent residential area access control system and other fields have shown more significant application results. With the rapid development of intelligent transportation system, the single pattern recognition technology based on radio frequency identification or video recognition has been unable to meet the increasingly complex road conditions for the accuracy and robustness of the vehicle identification system. Therefore, this paper mainly focuses on the anti-collision problem of tags and license plate location in RFID and video identification. The main research work is as follows: 1. Aiming at the low throughput of the existing binary search algorithms in radio frequency system, this paper improves the coding method of the existing binary search algorithms, and obtains an improved binary search optimization coding algorithm. Using the uniqueness of tag sequence number and Manchester coding, the algorithm can accurately identify the location of tag collision, and encode the characters sent by reader by binary optimization. The improved algorithm significantly reduces the number of searches and the length of the characters sent, which improves the efficiency of the RFID system. 2. Aiming at the interference of objective factors, such as different illumination angle and distance, vehicle moving rate, changeable, complex background and so on, on the aspect of contrast, this paper improves the existing wavelet transform algorithm. An improved license plate location algorithm based on wavelet transform is presented. The algorithm is suitable for the recognition of natural scene under the condition of angle change, distance change and illumination intensity change. Theoretical analysis and simulation results show that the proposed improved algorithm is feasible. The performance of binary search optimization coding algorithm is superior to the traditional binary search algorithm and dynamic binary search algorithm in terms of system throughput, transmission data length and search times. The license plate location algorithm based on wavelet transform has achieved good results when the shooting angle is more than 30 degrees and the shooting distance is more than 3 meters.
【学位授予单位】：兰州理工大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：U495;TP391.44

【相似文献】