面向图像去雾处理全局化增强方法的研究与应用
发布时间：2018-09-03 20:48
【摘要】：近年来,由于经济的快速发展和计算机科技的普遍应用,应用在户外的监控系统对图像处理技术的要求也越来越高。但是环境问题的日益突出给这些技术的应用带来了挑战,尤其是当前严重的雾、霾等恶劣天气。恶劣的天气不仅会影响人们的日常生活,给生活带来不便,并且使获得的图像的质量变低,最终成像效果变差。严重时更可能使某些户外的监控系统无法正常运行,而且在安全方面存在很大隐患,尤其是交通方面。是以,对雾霾天气下所拍摄的图像进行图像增强处理无论是在日常生活还是专门部门都变得非常重要而且意义重大。本文主要的研究工作如下:(1)从图像增强的理论基础、数字图像的表示、获取方法、处理过程和各种图形类型的基本特点等基础概念入手,综合比较了两种局部化增强图像方法,基于局部对比度方法和基于局部方差法。并提出了在小波域的基础上做改进的Retinex算法,通过与直方图均衡化的去雾效果做实验比较可得,改进的算法有着良好的去雾效果,且各种客观指标也比较高。(2)在频域变换的理论基础上,分析并使用Sym系列小波变换的图像处理增强方法,深入研究了小波变换技术在图像加强方面的方法和应用,阐述了小波分解时选取合适的小波基的重要性,以及分解后获得的不同高低频信息采取基于各自特点差别的处理机制,会使最终效果有很大不同,以此来使图像的清晰度得到改善。将该方法的去雾效果与同态滤波算法去雾做了比较,实验证明本文方法去雾效果更好。最后还将该方法运用到了交通车辆监控系统中,其处理效果也很好。
[Abstract]:In recent years, with the rapid development of economy and the widespread application of computer science and technology, the requirement of image processing technology for outdoor surveillance system is more and more high. However, the increasingly prominent environmental problems have brought challenges to the application of these technologies, especially the current severe fog, haze and other bad weather. The bad weather will not only affect people's daily life, bring inconvenience to life, but also make the quality of the image become lower, and the final imaging effect will become worse. It is more likely that some outdoor monitoring systems will not work properly when it is serious, and there are great hidden dangers in safety, especially in traffic. Therefore, image enhancement processing of images taken in haze weather has become very important and significant both in daily life and in specialized departments. The main work of this paper is as follows: (1) starting with the basic concepts of image enhancement, such as image enhancement theory, digital image representation, acquisition method, processing process and the basic characteristics of various graphic types, two localization enhancement methods are comprehensively compared. Based on local contrast method and local variance method. An improved Retinex algorithm based on wavelet domain is proposed. Compared with the experimental results of histogram equalization, the improved algorithm has a good de-fogging effect. And various objective indexes are also relatively high. (2) based on the theory of frequency domain transform, the image processing enhancement method of Sym series wavelet transform is analyzed and used, and the method and application of wavelet transform technology in image enhancement are deeply studied. The importance of selecting suitable wavelet bases in wavelet decomposition is expounded, and the processing mechanism based on the characteristics and differences of the different high and low frequency information obtained by wavelet decomposition will make the final effect very different. In order to improve the clarity of the image. Compared with the homomorphic filtering algorithm, the effect of the proposed method is better than that of the homomorphic filtering algorithm. Finally, the method is applied to the traffic vehicle monitoring system, and its processing effect is very good.
【学位授予单位】：山东师范大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP391.41
[Abstract]:In recent years, with the rapid development of economy and the widespread application of computer science and technology, the requirement of image processing technology for outdoor surveillance system is more and more high. However, the increasingly prominent environmental problems have brought challenges to the application of these technologies, especially the current severe fog, haze and other bad weather. The bad weather will not only affect people's daily life, bring inconvenience to life, but also make the quality of the image become lower, and the final imaging effect will become worse. It is more likely that some outdoor monitoring systems will not work properly when it is serious, and there are great hidden dangers in safety, especially in traffic. Therefore, image enhancement processing of images taken in haze weather has become very important and significant both in daily life and in specialized departments. The main work of this paper is as follows: (1) starting with the basic concepts of image enhancement, such as image enhancement theory, digital image representation, acquisition method, processing process and the basic characteristics of various graphic types, two localization enhancement methods are comprehensively compared. Based on local contrast method and local variance method. An improved Retinex algorithm based on wavelet domain is proposed. Compared with the experimental results of histogram equalization, the improved algorithm has a good de-fogging effect. And various objective indexes are also relatively high. (2) based on the theory of frequency domain transform, the image processing enhancement method of Sym series wavelet transform is analyzed and used, and the method and application of wavelet transform technology in image enhancement are deeply studied. The importance of selecting suitable wavelet bases in wavelet decomposition is expounded, and the processing mechanism based on the characteristics and differences of the different high and low frequency information obtained by wavelet decomposition will make the final effect very different. In order to improve the clarity of the image. Compared with the homomorphic filtering algorithm, the effect of the proposed method is better than that of the homomorphic filtering algorithm. Finally, the method is applied to the traffic vehicle monitoring system, and its processing effect is very good.
【学位授予单位】：山东师范大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP391.41
【参考文献】
相关期刊论文 前10条
1 孙晓晓;杨峰;;基于小波变换融合技术的去雾方法[J];山东师范大学学报(自然科学版);2016年02期
2 段世杰;黄华;王鹏飞;康永杰;;一种单帧图像的快速去雾方法[J];软件;2015年05期
3 郭奇;吕晓光;;基于小波变换的图像增强的实现研究[J];传感器世界;2015年03期
4 钱小燕;;单一图像多滤波联合快速去雾算法[J];科学技术与工程;2015年06期
5 李滚;吴R挤，
本文编号：2221123
本文链接：https://www.wllwen.com/kejilunwen/ruanjiangongchenglunwen/2221123.html
