人行横道交通灯识别算法设计及ARM实现
发布时间：2018-09-01 09:36
【摘要】：近年来,随着汽车的普及,人行横道通行安全和效率问题日益突出。为了探究降低交通事故率的有效方法,基于机器视觉的人行横道交通灯识别技术就成了目前的研究热点,同时也为盲人助行工具的研制提供一定的理论依据。 本文提出了一种在复杂场景下人行横道交通信号灯的检测与识别方法,该方法既能实现在复杂场景下检测人行横道交通信号灯,又能识别出信号灯的具体状态。具体的讲,本文在充分研究、归纳人行横道交通信号灯特征属性的基础上,首先建立了信号灯定位与识别系统的框架,并将系统框架分为图像获取、定位识别算法实现、ARM硬件实现三部分。图像获取又分为图像的采集与压缩,该框架结构简单、易于实现。基于数字图像处理的人行横道信号灯的识别与定位算法是本文研究的重点,针对现有交通灯检测方法仅适用于简单场景的不足,本文提出了一种基于HSI与RGB颜色空间相结合的阈值分割算法来定位与识别人行横道信号灯,并用Matlab编程实现该算法。现行大部分交通灯识别算法只能在PC上实现,而不能应用于嵌入式可移动设备上,针对此不足,本文提出并实现了C++程序实现该算法,并经过交叉编译后成功移植到搭载liunix操作系统的ARM9开发板,完成了算法的硬件实现。
[Abstract]:In recent years, with the popularity of cars, crosswalk safety and efficiency issues have become increasingly prominent. In order to explore an effective method to reduce traffic accident rate, the recognition technology of traffic lights in crosswalk based on machine vision has become a hot research topic at present, and it also provides a certain theoretical basis for the development of walking aid tools for blind people. This paper presents a detection and recognition method for crosswalk traffic lights in complex scenes. This method can not only detect crosswalk traffic lights in complex scenes, but also recognize the specific state of traffic lights. Specifically, based on the full study and induction of the characteristic attributes of traffic lights in crosswalk, this paper first establishes the frame of the signal light location and recognition system, and divides the system frame into image acquisition. The location recognition algorithm realizes three parts of arm hardware. Image acquisition is divided into image acquisition and compression. The framework is simple and easy to implement. The recognition and location algorithm of crosswalk signal based on digital image processing is the focus of this paper. The existing traffic light detection method is only suitable for simple scene. In this paper, a threshold segmentation algorithm based on HSI and RGB color space is proposed to locate and recognize crosswalk signal lights. The algorithm is implemented by Matlab programming. Most of the current traffic light recognition algorithms can only be implemented on PC, but not on embedded mobile devices. In view of this deficiency, this paper proposes and implements the C program to implement the algorithm. After cross-compiling, the algorithm was successfully transplanted to the ARM9 development board with liunix operating system, and the hardware implementation of the algorithm was completed.
【学位授予单位】：内蒙古大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：U491.54
本文编号：2216792
[Abstract]:In recent years, with the popularity of cars, crosswalk safety and efficiency issues have become increasingly prominent. In order to explore an effective method to reduce traffic accident rate, the recognition technology of traffic lights in crosswalk based on machine vision has become a hot research topic at present, and it also provides a certain theoretical basis for the development of walking aid tools for blind people. This paper presents a detection and recognition method for crosswalk traffic lights in complex scenes. This method can not only detect crosswalk traffic lights in complex scenes, but also recognize the specific state of traffic lights. Specifically, based on the full study and induction of the characteristic attributes of traffic lights in crosswalk, this paper first establishes the frame of the signal light location and recognition system, and divides the system frame into image acquisition. The location recognition algorithm realizes three parts of arm hardware. Image acquisition is divided into image acquisition and compression. The framework is simple and easy to implement. The recognition and location algorithm of crosswalk signal based on digital image processing is the focus of this paper. The existing traffic light detection method is only suitable for simple scene. In this paper, a threshold segmentation algorithm based on HSI and RGB color space is proposed to locate and recognize crosswalk signal lights. The algorithm is implemented by Matlab programming. Most of the current traffic light recognition algorithms can only be implemented on PC, but not on embedded mobile devices. In view of this deficiency, this paper proposes and implements the C program to implement the algorithm. After cross-compiling, the algorithm was successfully transplanted to the ARM9 development board with liunix operating system, and the hardware implementation of the algorithm was completed.
【学位授予单位】：内蒙古大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：U491.54
【参考文献】
相关期刊论文 前2条
1 徐成;谭乃强;刘彦;;基于Lab色彩空间和模板匹配的实时交通灯识别算法[J];计算机应用;2010年05期
2 程开富;;CMOS图像传感器的基本原理及设计考虑[J];电子元器件应用;2002年08期
相关硕士学位论文 前1条
1 李天琦;CMOS图像传感器像素光敏器件研究[D];哈尔滨工程大学;2013年
，本文编号：2216792
本文链接：https://www.wllwen.com/kejilunwen/jiaotonggongchenglunwen/2216792.html
