一种基于对称性边缘的多光谱图像配准方法

Multispectral Image Registration Based on Symmetrical Edge

摘要:

多光谱成像技术是遥感领域的一项重要技术。多光谱图像配准技术可以提高遥感图像的应用效率和能力。基于边缘结构在多光谱图像中较为稳定这一特性,提出了一种基于对称性边缘的多光谱图像配准方法,该方法主要包含“图像比例调整”、“对称性边缘提取”和“互信息配准”等三个步骤。基于天宫二号宽波段成像仪的遥感数据设计了一种多子图配准方案,验证了所提方法在可见光、短波红外和长波红外三个谱段之间图像配准的有效性。设置对比实验,将所提配准方法与其他多光谱图像配准方法进行比较,结果表明该方法在天宫二号宽波段遥感图像的配准中具有较高的精度。

Abstract:

Multispectral imaging technology is an important technology in the field of remote sensing. Multispectral image registration technology can improve the application efficiency and ability of remote sensing image. Based on the stability of edge structure in multispectral images, a new registration method for multispectral images based on symmetrical edge was proposed in this paper, which mainly includes three steps of image scale adjustment, symmetrical edge extraction and mutual information registration. In this paper, based on the remote sensing data of Tiangong-2 wide-band imaging spectrometer, a multi-subimage process was carried out to verify the effectiveness of the proposed method in image registration between bands of visible and near-infrared, short-wave infrared and long-wave infrared. Contrast experiments indicate that the proposed method has higher accuracy in the wide-band remote sensing image registration of Tiangong-2.