基于DM6437的不同照度下阈值选取对车道偏离预警系统影响的研究

发布时间：2018-01-18 03:13

本文关键词：基于DM6437的不同照度下阈值选取对车道偏离预警系统影响的研究　出处：《长安大学》2014年硕士论文　论文类型：学位论文

【摘要】：车道偏离预警系统在现在道路交通安全中起着越来越重要的作用，是现代交通中重要的安全辅助系统。车道偏离预警系统中最关键的一步就是对车道线的识别和提取，，其中阈值的选取对车道线的识别有着重要的影响。本文通过试验分析了照度对二值化阈值选取的影响。本文的硬件平台是TMS320DM6437评估板，以此作为车道偏离预警系统的核心，并根据板载Flash的特点，编写相应的烧写Flash的程序，将程序固化在评估板的Norflash中，实现了系统的脱机运行。为了研究照度对阈值选取的影响，本文为硬件平台添加了照度传感器模块；照度传感器通过I2C总线与51单片机通讯，51单片机将采集到的照度数据通过串口通讯实时地将照度发送给DM6437，并且照度值能够通过1602液晶显示屏实时显示出来。本文对不同照度下的道路图像进行对比度增强，并重点研究了照度对二值化阈值选择的影响。通过试验，分析不同照度下，阈值的具体选取过程，研究了照度与阈值之间的规律关系。根据二值化道路图像的特点，选用合适的车道线识别算法，成功提取车道线。并在原有偏离报警算法的基础上，加入了车道线夹角判别方法，提高了车道偏离报警的及时性。通过实车试验，在不同照度条件下，对车道偏离预警系统做了大量试验，验证系统算法，并将照度划分成不同的区间，区分强照度、一般照度和低照度，总结了照度与二值化阈值之间的规律，进而分析了不同照度下阈值的选取对车道线的识别产生的影响。
[Abstract]:Lane deviation warning system plays an increasingly important role in road traffic safety. It is an important safety assistant system in modern traffic. The most important step of lane deviation warning system is to identify and extract lane line. The selection of threshold value has an important effect on lane line recognition. This paper analyzes the influence of illumination on binarization threshold selection through experiments. The hardware platform of this paper is TMS320DM6437 evaluation board, which is the core of driveway deviation warning system, and according to the characteristics of Flash on board. In order to study the influence of illuminance on threshold selection, the corresponding program of burning and writing Flash is written, and the program is solidified in the Norflash of the evaluation board to realize the off-line operation of the system. In this paper, the illuminance sensor module is added to the hardware platform. The illuminance sensor communicates with 51 single chip microcomputer through I2C bus. The illumination data collected by single chip computer can be transmitted to DM6437 through serial communication in real time. And the illuminance value can be displayed by 1602 LCD in real time. In this paper, the contrast of road images under different illuminance is enhanced, and the influence of illumination on binarization threshold selection is studied. Through experiments, the specific selection process of threshold is analyzed under different illuminance. The relationship between illumination and threshold is studied. According to the characteristics of binary road image, the appropriate lane line recognition algorithm is selected, and the lane line is extracted successfully, and based on the original deviation alarm algorithm. The method of lane line angle discrimination is added to improve the timeliness of lane deviation alarm. Through the actual vehicle test, under different illumination conditions, a large number of experiments on lane deviation warning system, verify the system algorithm, and will be divided into different sections of illuminance, distinguish strong illumination, general illumination and low illuminance. The law between illuminance and binarization threshold is summarized, and the influence of the selection of threshold on lane line recognition is analyzed.
【学位授予单位】：长安大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：U463.6;U495

【参考文献】