摘要智能驾驶辅助技术能自动感知影响车辆行驶过程中的一些道路状况及突发事件，提高车辆行驶的安全性，在军事及国民经济诸多领域存在广泛应用。红外双目立体视觉，可在夜间或者会车遭遇对向强光照射条件下获得视场清晰的立体视觉，并可通过算法计算得目标的被动测距数值，对比其他成像方式有无与伦比的优势。80519

本文首先阐述了双目立体视觉的机理和研究方法，研究了双目立体相机标定和基于红外显著目标特征提取的立体匹配方法。为解决常规算法匹配精度高但效率不高的现状，本文采用的是SUSAN算子，利用该算法对红外立体显著目标进行特征提取和区域匹配，之后计算得出相应的视差值，利用视差图进行插值从而最终得到稠密的视差图。根据双目立体视觉测距原理，我们利用视差图获取了距离信息，通过构建了立体匹配测距系统，对不同距离下目标进行了测距实验。

结果表明，本文研究的立体匹配算法具有较好的显著目标测距精度，同时也具有提升平衡匹配准确率和计算效率的优点。

毕业论文关键词  智能驾驶辅助，红外双目立体，立体匹配，双目测距。

毕业设计说明书外文摘要

Title  The research of infrared stereo intelligent driving auxiliary devices ranging technology 

Abstract The technology of intelligent driving assistance can automatically sense the road conditions and emergencies during driving the car and improve vehicle safety, Meanwhile this technology is widely used in the presence of the military and the national economy。 IR binocular stereo vision will assist the car to obtain clear field of vision when it be at night or facing under bright light conditions 。 and can be calculated by the target passive ranging algorithm value。 compared to other imaging modalities, it have unparalleled advantages。

This paper describes the mechanism of binocular stereo vision and research methods, the binocular stereo camera calibration and stereo matching method based on infrared significant target feature extraction。 To solve the high-precision conventional matching algorithm but the efficiency is not high status, we use SUSAN operator is using this algorithm significantly infrared target feature extraction and stereo matching area and then calculated the corresponding disparity value using a disparity map interpolating ultimately obtain dense disparity map。 According to a significant target extraction and stereo matching based on the above, we constructed a stereo matching ranging system, was the target of different distances ranging experiments。

The results show that the study of stereo matching algorithm has better range accuracy remarkable goals, but also has the benefits of increased accuracy and balance match the computational efficiency。

Key words  Intelligent driving assistance  Infrared binocular  stereo stereo  matching binocular  distance measurement 

目   次

1  绪论 1

1。1  研究背景及意义 1

1。2  红外波段下立体匹配测距国外研究现状 2

1。3  如何获取显著目标的距离信息 2

1。4  本文的研究思路和结构安排 4

2  双目视觉系统的相机标定 6

2。1  理论基础 6

2。2  平面标定法 8

2。3  极线校正的原理