多摄像头全景图像拼接的实时视频处理技术研究

发布时间：2018-07-17 01:31

【摘要】：现如今,随着数码技术和信息技术的飞速发展,人们对图像采集设备倍受关注,同时也对图像的分辨率和场景图像不断有更高的要求。单凭一个摄像机可能无法满足消费者的需求,毕竟单个摄像机所能够拍摄的范围有限,若想获得高分辨率的场景图像,单个摄像机只能通过调节焦距的方式来实现,但问题出现了,这样会导致图片分辨率降低,图片变模糊,给人造成不良的视觉效果。因此,可以通过多个摄像头来采集图像,并完成全景图像的拼接。实时全景监控系统则是多摄像头实时拼接技术主要应用之一,它很好解决了传统监控系统存在多视角场景监控,相对凌乱的问题,在安防监控市场领域(通信、商店超市、道路码头等)有不错的表现。本文主要对多摄像头全景图像拼接的实时视频处理技术进行研究,并介绍多摄像头全景图像实时拼接系统架构。首先,简要介绍多摄像头全景图像拼接的实时视频处理技术研究背景和研究现状,以及与多摄像头全景图像拼接相关的基础理论知识。然后,提出多摄像头全景图像实时拼接系统整体方案和架构,对其中各个部分的功能进行简要描述。接着,针对多摄像头全景图像拼接中两个关键环节-图像配准和图像融合,进行详细的介绍。列举出图像配准和图像融合常用的方法,通过实验对比选择合适的方法来完成全景图像拼接。图像配准部分通过分析SIFT特征配准算法原理及实现,图像融合部分则采用多分辨率融合方法,确保图像融合质量。同时,考虑到多个摄像头附近的光照强度不同,分析颜色不同带来的影响,提出图像颜色校正方法,完成多摄像头全景图像颜色校正,保证全景图的效果。最后,通过对多摄像头全景实时图像拼接系统进行测试验证,测试结果表明本文设计的方案能较好地实现全景图像的拼接,对今后多摄像头全景图像实时拼接系统的开发和研究提供了一定的参考价值。
[Abstract]:Nowadays, with the rapid development of digital technology and information technology, people pay more attention to image acquisition equipment, but also have higher requirements for image resolution and scene image. A single camera may not be able to meet the needs of consumers. After all, a single camera can only be achieved by adjusting the focal length to obtain a high-resolution scene image. However, problems arise, which can lead to lower resolution and blur images, resulting in poor visual effects. Therefore, images can be captured by multiple cameras and the panoramic images can be stitched together. Real-time panoramic monitoring system is one of the main applications of multi-camera real-time splicing technology. It solves the problem of multi-view scene monitoring and relatively messy in traditional monitoring system. It is in the field of security monitoring market (communication, store supermarket, store supermarket). Road, wharf, etc.) have a good performance. In this paper, the real-time video processing technology of multi-camera panoramic image mosaic is studied, and the architecture of multi-camera panoramic image real-time mosaic system is introduced. Firstly, the background and research status of real-time video processing technology for multi-camera panoramic image mosaic are briefly introduced, as well as the basic theoretical knowledge related to multi-camera panoramic image mosaic. Then, the overall scheme and architecture of real-time multi-camera panoramic image mosaic system are proposed, and the functions of each part are briefly described. Then, two key links in multi-camera panoramic image mosaic, image registration and image fusion, are introduced in detail. The common methods of image registration and image fusion are listed. The principle and implementation of sift feature registration algorithm are analyzed in the image registration part, and the multi-resolution fusion method is used in the image fusion part to ensure the image fusion quality. At the same time, considering the different illumination intensity near several cameras, the influence of different colors is analyzed, and a color correction method is proposed to complete the color correction of multi-camera panoramic images to ensure the effect of panoramic images. Finally, by testing the multi-camera panoramic real-time image mosaic system, the test results show that the scheme designed in this paper can better achieve panoramic image mosaic. It provides some reference value for the development and research of multi-camera panoramic image mosaic system in the future.
【学位授予单位】：华侨大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP391.41

【相似文献】