夜晚铁路图像增强技术的研究

发布时间：2018-03-27 23:41

本文选题：低照度　切入点：铁路　出处：《北京交通大学》2017年硕士论文

【摘要】：随着城市规模的不断发展,铁路基础设施陆续建设完成,列车技术取得了重大进步,同时也对列车运行安全提出了更高的要求。物体非法入侵铁路轨间区域,引发铁路交通事故,对铁路的运输效率以及人民生命安全造成了极大威胁。在铁路运营中,通常需要使用铁路眺望、图像识别等方式探测轨间异物。由于低照度环境下采集的图像对比度低、视觉效果差,细节信息无法清晰呈现出来,这使得后续的图像特征提取、分析与识别受到限制。因此,对夜晚铁路图像进行增强处理具有较强的实用性。论文分析了基于非物理模型和物理模型的低照度图像增强算法,阐述了低照度图像增强算法的基本过程并分析了其原理,结合夜晚铁路环境,指出了现有"基于暗通道的低照度图像增强算法"与传统增强算法相比具有更高的保真度,对细节的保留效果更好。将现有"基于暗通道的低照度图像增强算法"应用于铁路图像中,出现曝光率过高以及强光源区域失效等问题,结合铁路图像特性从原理以及实验的角度对问题进行剖析。本文主要研究的问题如下:第一,针对现有"基于暗通道的低照度图像增强算法"在铁路图像强光源区域失效原因进行分析,通过设定透射率阈值将强光源区域进行图像分割,对该区域使用局部方差对其进行增强处理。第二,验证暗通道先验理论在轨间区域的适用性,并提出了适用于铁路图像的暗通道理论。根据铁轨图像的亮度和斜率特征,结合边缘提取算法和Hough变化算法对轨间区域进行截取,基于铁路图像的暗通道理论对夜晚轨间区域进行增强处理。实验结果表明,本文算法可以有效提升夜晚铁路图像对比度,突出细节信息,从而获得更加清晰、自然的高质量图像。
[Abstract]:With the continuous development of the city scale, the construction of railway infrastructure has been completed one after another, and the train technology has made great progress. At the same time, it has put forward higher requirements for the safety of train operation. Causes of railway traffic accidents, which pose a great threat to the efficiency of railway transportation and the safety of people's lives. In the operation of railways, it is usually necessary to use railways to look out. Image recognition and other methods are used to detect foreign bodies between tracks. Because of low contrast, poor visual effect and poor visual effect, the detailed information can not be clearly presented, which makes the subsequent image feature extraction, analysis and recognition limited. It is very practical to enhance railway images at night. This paper analyzes the low illuminance image enhancement algorithm based on non-physical model and physical model, expounds the basic process of low illuminance image enhancement algorithm and analyzes its principle. Combined with the night railway environment, it is pointed out that the existing "low illuminance image enhancement algorithm based on dark channel" has a higher fidelity than the traditional enhancement algorithm. The existing "low illuminance image enhancement algorithm based on dark channel" is applied to railway image, which has the problems of high exposure rate and failure of strong light source area, etc. Combined with the characteristics of railway images, the problems are analyzed from the angle of principle and experiment. The main problems of this paper are as follows: first, This paper analyzes the causes of the failure of the existing "low illuminance image enhancement algorithm based on dark channel" in the railway image strong light source region, and segments the strong light source region by setting the transmittance threshold. The local variance is used to enhance the region. Secondly, the applicability of the dark channel priori theory in the interorbital region is verified, and the dark channel theory suitable for railway image is proposed. According to the brightness and slope characteristics of the railway image, Combined with edge extraction algorithm and Hough variation algorithm, the interorbit region is intercepted and enhanced based on dark channel theory of railway image. The experimental results show that the proposed algorithm can effectively improve the contrast of the night railway image. Highlight the details to get a clearer, natural, high-quality image.
【学位授予单位】：北京交通大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：U298

【参考文献】