基于智能终端的车辆碰撞预警系统的设计与实现

发布时间：2018-10-11 10:24

【摘要】：随着社会经济快速发展和人们生活水平的提高,城市道路上几乎到处是形形色色的车辆,方便生活的同时,给人们出行带来了极大的隐患,交通事故的频繁发生。为预防交通事故的发生,目前汽车辅助驾驶系统已经成为智能交通领域的研究热点,其中,车辆防撞预警是重要的组成部分。与此同时,移动互联网发展的大趋势下,基于Android平台的智能手机使用最为广泛,智能手机集成了高分辨率摄像头、GPS、加速度传感器使得它成为了汽车辅助驾驶最好的平台。本文主要研究基于Android平台的车辆碰撞预警系统的关键技术,包括车辆检测技术、车辆间的距离测量和防撞预警技术,具体的研究内容如下:(1)研究并实现车辆检测算法,基于感兴趣区域对采集的道路图像进行预处理,提取车辆底部阴影特征,根据阴影底部边缘的位置生成车辆假设区域。(2)对大量车辆的正样本和负样本提取Haar-like特征使用AdaBoost分类器进行训练,对车辆假设区域提取Haar-like特征,并使用训练好的模型进行分类,验证是否为车辆,最终确定车辆的位置。使用车底阴影特征和Haar-like特征相结合进行图像中车辆的检测。(3)根据检测到的车辆使用单目视觉测距的方法进行车辆间距离测量,设置安全距离,当车距小于安全距离时发出警报提醒驾驶人员,并设计实验进行车距测量。(4)设计并实现了基于Android的车辆防撞预警客户端软件,包括图像采集模块、图像预处理模块、车辆检测模块、车距检测模块和防撞预警模块,利用智能手机的摄像头进行道路图像采集并对采集到的图像进行预处理、车辆检测、车距测量,当车距小于安全距离时由智能手机发出警报提醒驾驶人员。
[Abstract]:With the rapid development of social economy and the improvement of people's living standard, there are almost all kinds of vehicles on the city road. In order to prevent traffic accidents, the automobile assistant driving system has become the research hotspot in the field of intelligent transportation, among which, the vehicle anti-collision warning is an important part. At the same time, with the development of mobile Internet, smart phones based on Android platform are most widely used. Smart phones integrate high-resolution camera, and GPS, accelerometer makes it the best platform for automobile assisted driving. This paper mainly studies the key technologies of vehicle collision warning system based on Android platform, including vehicle detection technology, distance measurement and anti-collision warning technology. The specific research contents are as follows: (1) Research and implement vehicle detection algorithm. Based on the region of interest, the acquired road image is preprocessed to extract the shadow features of the bottom of the vehicle. The vehicle hypothesis region is generated according to the position of the shadow bottom edge. (2) Haar-like feature extraction from a large number of positive and negative samples is trained by AdaBoost classifier, and Haar-like feature is extracted from the vehicle hypothesis region, and the trained model is used to classify the vehicle hypothesis region. Verify that it is a vehicle, and finally determine the location of the vehicle. The vehicle detection in the image is carried out by combining the shadow feature of the vehicle bottom with the Haar-like feature. (3) the distance between the vehicles is measured by the method of monocular vision ranging, and the safe distance is set up according to the detection method. When the distance is less than the safe distance, the alarm is issued to the driver, and the experiment is designed to measure the distance. (4) the vehicle anti-collision warning client software based on Android is designed and implemented, including the image acquisition module and the image preprocessing module. Vehicle detection module, vehicle distance detection module and collision warning module, using the camera of smart phone to collect road image and preprocess the collected image, vehicle detection, vehicle distance measurement, The smartphone alerts drivers when the distance is less than a safe distance.
【学位授予单位】：南京理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：U463.6;TP391.41

【相似文献】