节点文献
数字交通图像融合算法研究
Research on Digital Image Fusion Algorithm for Traffic
【作者】 王浩宇;
【导师】 周栩;
【作者基本信息】 吉林大学 , 软件工程(专业学位), 2016, 硕士
【摘要】 随着国民经济的飞速发展,车辆交通已经成为了众人关注的焦点,城市交通系统的流畅性,关系到市民生活幸福指数。国内外交通部门及相关学者在不断的升级研究各自交通控制系统,为了改善智能交通系统的服务指标,大多数智能交通系统的发展依赖于其硬件的改善,例如提高摄像传感器的采集质量,增加各项传感功能等,耗费大量物力财力,基于以上措施,直接对交通图像进行数字处理,其成本更低,效率更高。对单个传感器所采集的图像进行处理已经远远不能满足智能交通系统发展,多传感器智能交通图像信息融合的优势就越发明显了。本文的主要研究方向就是交通多传感器数字信号融合理论和算法,提出了小波变换与HSI变换相结合的适合于交通图像的融合算法,又针对于交通目标车辆提出了基于灰度近似的交通图像融合算法,通过MATLAB实验平台进行目标仿真,效果明显。论文第一章首先介绍了当今世界城市智能交通控制系统的发展情况,通过了解智能交通的发展引出本文所要研究的智能交通系统中的视频模块的多传感器融合技术的融合理论以及发展概况,并详细解释了图片的评价标准。接下来是本篇论文的重点,论文在第三章提出了小波变换结合HSI变换的融合算法,融合结果图像交通目标特征突出明显,改善了单一交通传感器的局限性。最后提出了基于灰度近似的交通图像融合算法,此算法通过构造一个新的窗函数,其像元的选择通过设定中心点来选取,近景图像信息的保留原则是通过像元与中心点的距离来判定,依据像元与中心点的远近来选择保留信息的程度,这样可以消弱图像融合的边缘痕迹,效果明显。
【Abstract】 With the rapid development of the national economy, the traffic has become the focus of attention of the public, the fluency of the urban traffic system, related to the life of the people happiness index. Domestic and foreign transportation departments and relevant scholars in the continuous upgrading of research to respective traffic control system, in order to improve the service index of intelligent transportation system(ITS), most of the intelligent transportation system development depends on the improvement of the hardware,such as raising the camera sensor acquisition quality, increase the sensing function, cost a lot of money and, on the basis of the above measures directly the traffic image digital processing, the lower cost and higher efficiency.The image of a single sensor processing can not meet the development of intelligent transportation system, the advantages of multi sensor intelligent traffic image information fusion is more obvious.In this thesis, the main research direction is traffic multi digital sensor signal fusion theory and algorithm,proposed the combination of wavelet transform and HSI transform for fusion algorithm of the traffic image, with needle for vehicle traffic objects put forward traffic images based on the approximate gray fusion algorithm, through MATLAB experiment platform for target simulation, the effect is obvious.The first chapter of this thesis first introduces the nowadays world urban intelligent traffic control system development, through understanding the development of intelligent transportation leads this to video module of intelligent transportation system in multi sensor fusion technology integration theory and the development situation. And a detailed explanation of the image evaluation standards.Next is the focus of this paper, paper in the third chapter the HSI transform fusion algorithms based on wavelet transform and, fusion algorithm is effective in improving the quality of traffic image, fusion results traffic image target feature prominent, improve the limitations of single traffic sensors. Finally, the paper puts forward the fusion algorithm based on approximate gray image of the traffic, the algorithm by constructing a new window function, the choice of the pixel through set the center point to select, close range image information of the reservation principle is through the pixel and the center point of the distance to determine, according to the distance of the pixels and center point to selectthe information lost. This can weaken the image fusion edge traces of, the effect is obvious.
【Key words】 Traffic image; HSI transform; The wavelet transform; Gray approximate;
- 【网络出版投稿人】 吉林大学 【网络出版年期】2016年 09期
- 【分类号】TP391.41
- 【下载频次】123