基于纹理公因子的异源图像配准方法

发布时间：2018-03-22 13:25

本文选题：图像配准　切入点：异源图像　出处：《计算机工程》2016年11期 　论文类型：期刊论文

【摘要】：异源图像的成像机理不同,导致同一场景的异源图像之间存在着较大的形变或者光照等差异,异源图像配准很难直接运用尺度不变特征变换算法(SIFT)和快速鲁棒特征算法(SURF)等方法。为此,针对异源图像提出一种基于纹理公因子的配准方法。根据傅里叶变换将异源图像变换到频率域并用Gabor模版进行滤波处理,在空间域中利用Sobel算子对异源图像进行纹理公因子提取,采用自定义规则选择匹配点,通过随机抽样一致性算法对匹配点对进行提纯,根据提纯后的匹配点对求解单应变换参数,经过坐标变换及插值分析实现异源图像配准。实验结果表明,与SIFT和SURF等算法相比,该方法的匹配准确率较高,鲁棒性和异源图像配准效果较好。
[Abstract]:The imaging mechanism of the heterologous image is different, which leads to the difference of deformation or illumination between the different images of the same scene. It is difficult to directly use scale invariant feature transform algorithm (sift) and fast robust feature algorithm (SURF) in heterogeneous image registration. A registration method based on texture common factor is proposed for heterogeneous images. According to Fourier transform, the heterogeneous images are transformed into frequency domain and filtered by Gabor template. In spatial domain, Sobel operator is used to extract the common factors of different images, and the matching points are selected by custom rules, and the matching points are purified by random sampling consistency algorithm. According to the extracted matching point pairs, the parameters of monogram transformation are solved, and the heterologous image registration is realized by coordinate transformation and interpolation analysis. The experimental results show that the matching accuracy of this method is higher than that of SIFT and SURF algorithms. Robustness and heterologous image registration are better.
【作者单位】：长沙理工大学物理与电子科学学院;国防科学技术大学航天科学与工程学院;国防科学技术大学空天图像测量与视觉导航研究中心;长沙亿旭机电科技有限公司;
【基金】：湖南省教育厅科学研究项目“远程实验控制关键技术研究”(15K009)
【分类号】：TP391.41

【相似文献】