图像配准关键技术综述

图像采集设备的快速发展带来了图像种类和数量的迅速增加,图像拼接、融合等视觉系统提高了对图像配准的技术需求。图像配准利用原始图像数据或图像特征信息对输入图像的重叠区域进行变换和对齐,是为后续图像拼接、图像融合等图像处理任务提供操作基准的核心技术,是当前计算机视觉领域的热点研究问题。首先介绍了图像配准技术的主要应用领域及重大发展前景,然后说明了图像配准技术的基本框架和技术难点,详细阐述了基于特征以及基于区域的图像配准技术的实现算法及关键技术,分析了不同配准算法的优缺点及适用性,并对图像配准技术的发展前景进行了讨论和展望。

Review of Key Technologies of Image Registration

The rapid development of image acquisition equipment has brought about a rapid increase in the number and type of images. Vision systems such as image stitching and fusion have increased the technical requirements for image registration. Image registration technology uses original image data or image feature information to transform and align the overlapping regions of the input images, which is the core technology for providing operational benchmarks for the follow-up image processing tasks such as image stitching and image fusion. It is a hot research topic in the current computer vision field. Firstly, the main application areas and major development prospects of image registration technology are introduced. Then, the basic framework and technical difficulties of image registration technology are explained. The implementation algorithms and key technologies of feature-based and region-based image registration technology are summarized. The advantages, disadvantages and applicability of different registration algorithms are analyzed. The development prospects of image registration technology are discussed and prospected.