摘要近年来，我国的公路交通行业发展迅速，发达的现代交通给人们的生活带来 了很大的便利，但同时交通拥挤、交通安全等问题也日渐突出。为了应对这些问 题，智能交通系统的概念被提出。它通过摄像头采集交通图像，之后将图像送入 CPU 进行处理，最后将结果反馈给驾驶员，达到辅助驾驶员判断，增加道路交通 安全性的目的。
本文对交通标志的检测和识别进行了如下研究：83852
（1）提取交通标志的颜色特征。将采集到的图像进行预处理后转换到 HSV 颜色空间中，提取出指定颜色的像素点。
（2）根据提取的颜色特征精确分割交通标志。对图像二值化后进行去噪、 平滑、形体学处理等操作，将交通标志完整地提取出来。
（3）对交通标志分类所用的 BP 神经网络算法做了研究，包括理论原理、实 际应用、代码实现等，并对其优缺点进行了总结和探究。
（4）对系统的功能进行验证。系统在 Windows 环境下开发，完成后将其移 植到 Linux 平台下进一步验证，研究在低成本硬件平台下系统的功能可靠性、工 作效率、实际可行性等问题。
基于以上的分析和研究，系统在树莓派 2B 硬件和 OpenCV 库的环境下得到了 验证，结果表明此系统确实可以实现交通标志的自主判别，成本较低，工作效率 较高，满足设计需求。
毕业论文关键词：路标识别；BP 神经网络；Linux；OpenCV
Abstract In recent years, road transport industry is developing rapidly in China。 Developed modern transportation has brought great convenience, but traffic congestion, traffic safety and other issues become more prominent。 To solve these problems, the concept of intelligent transportation systems have been proposed。 It collects traffic images through the camera, sends the images into CPU for processing, and final results will be fed back to the driver to assist driver judgment, increase road traffic safety。
In this paper, traffic sign detection and identification of research carried out as follows:
(1)Extracting color features of traffic signs。 Collected images will be preprocessing conversion to HSV color space to extract the pixels of specified color。
(2)Based on the extracted color feature accurate segmentation traffic signs。 After image binarization de-noising, smoothing, physical science of processing operations, a traffic sign completely extracted。
(3)BP neural network algorithm to classify traffic signs used to do the research, including theoretical principles, practical application, code, etc。, and their advantages and disadvantages are summarized and explored。
(4)Verifying system functionality。 System completed in the Windows environment, then transplant to Linux to further validation, Research on low-cost hardware platform to function in the system's reliability, efficiency and practical feasibility and other issues。
Based on the above analysis and research systems in support Raspberry Pi 2B hardware environment and OpenCV library has been verified, the results indicate that this system can really achieve self-determination of traffic signs, low cost, high efficiency, satisfied the design requirements 。
Keywords: traffic symbol detection；BP neural network；Linux；OpenCV
目 录
第一章 绪论
1。1 研究背景 1
1。2 路标识别系统的发展与研究现状 1
1。3 交通标志检测技术 2
1。4 本文主要内容 3
1。5 论文结构与内容安排 3 基于OpenCV的路标识别系统设计:http://www.youerw.com/zidonghua/lunwen_99070.html