基于暗通道优先对交通图像的去雾处理(论文12000字)
摘要：随着我国经济的高速发展，计算机对数字图像的处理技术在日常生活中被广泛运用，如道路监控、交通监控等。但经济发展也伴随着一系列的环境问题，首当其中的就是雾霾。雾霾天气的出现，不仅危害着人们的身体健康，同时也给出行带来许多的不便，雾霾天气下交通事故发生率大大提高，然而监控系统所拍摄到的图像都因为雾霾的存在，有着或多或少的失真，无法辨认车辆的细节信息，给处理交通事故带来不便。因此，对有雾图像做去雾处理得到有着清晰细节的图像是很有必要的。
图像去雾算法主要分为两大类：图像增强和图像复原。图像增强主要从图像的对比度与亮度入手，使图像看起来清晰；图像复原是建立图像退化的物理模型，通过数学变换求出模型中清晰图像的部分。本文的研究方向是图像复原，以大气散射模型和暗通道优先理论为中心，从以下三个方面展开本文的工作。
1）通过研究大气散射模型，分析雾霾天气下图像降质的原因，给出暗通道优先的经典理论并进行matlab仿真实验，发现其有所不足。
2）分析经典暗通道优先去雾算法在色彩和效率上有所欠缺，根据两点不足进行改进。通过修正天空透射率可以减少明亮部分的失真率，引入导向滤波可以提高去雾速度，图像的质量也有所提高。
3）将经典暗通道优先算法和改进后的两种算法以及同组合作同学所研究的同态滤波去雾法进行比较，从主客观两个方面分析各个算法的优劣。
关键词：暗通道优先；大气散射模型；交通图像去雾；导向滤波；
Traffic image defogging based on dark channel priority
Abstract:With the rapid development of China's economy, computer processing techniques for digital images are widely used in daily life, such as road monitoring and traffic monitoring. But economic development is also accompanied by a series of environmental problems, the first of which is smog. The appearance of smog weather not only endangers people's health, but also gives many inconveniences. The incidence of traffic accidents in foggy weather is greatly improved. However, the images captured by the surveillance system are due to the presence of smog. With more or less distortion, it is impossible to identify the details of the vehicle, which is inconvenient for handling traffic accidents. Therefor
e, it is necessary to defogg the foggy image to obtain images with clear details.
Image defogging algorithms fall into two main categories: image enhancement and image restoration. Image enhancement mainly starts from the contrast and brightness of the image to make the image look clear. Image restoration is a physical model to establish image degradation, and the part of clear image in the model is obtained by mathematical transformation.. The research direction of this paper is image restoration. Focusing on the atmospheric scattering model and the dark channel priority theory, this paper starts from the following three aspects.
1）By studying the atmospheric scattering model, the causes of image degradation under haze weather are analyzed, the classical theory of dark channel priority is given and MATLAB simulation experiments are carried out, and it is found that it is insufficient.
2）Analysis of the classic dark channel priority dehazing algorithm is lacking in color and efficiency, and is improved according to two shortcomings. By correcting the sky transmittance, the distortion rate of the bright portion can be reduced, and the introduction of the guide filter can improve the defogging speed, and the image quality is also improved.
3）The classic dark channel first algorithm and the improved two algorithms and the homomorphic filtering dehazing method studied by the same group of students are compared, and the merits and demerits of each algorithm are analyzed from the subjective and objective aspects.
Key words：Dark channel priority；Atmospheric scattering model; Traffic image defogging; Guided filter;
目 录
1 绪论 1
1.1研究的背景与意义 1
1.2国内外研究现状 1
1.2.1图像增强的去雾算法研究现状 1
1.2.2图像复原的去雾算法研究现状 2
1.3本文的主要工作 2
2 图像去雾概述 3
2.1有雾图像的特征介绍 3
2.2大气散射物理模型 4
2.3图像质量评价 4
2.3.2客观质量评价 5
3 基于暗通道优先的图像去雾 7
3.1暗通道优先理论 7
3.1.1暗原色 7
3.1.2透射率的粗略求解 7
3.1.3全局大气光A的估算 9
3.1.4求解清晰图像 9
3.2经典暗通道的不足 9
3.2.1颜色失真 9
3.3经典暗通道的改进 10
3.3.1天空透射率的修正 10
3.3.2导向滤波取代软抠图 11
4 算法应用于交通图像的仿真与对比 12
4.1基于同态滤波的图像去雾 12
4.2图像仿真 12
4.2.1仿真环境 12
4.2.2图像仿真 13
4.3质量分析 15
4.3.1主观质量分析 15
4.3.2客观质量分析 15
5.总结与展望 17
参考文献 18
致谢 19 |
