基于相位一致性的遥感图像匹配方法

Remote sensing image matching algorithm based on phase congruency

图像匹配的优劣是影响图像拼接质量的一个重要因素,而传统的图像匹配方法对辐射畸变敏感.为了解决这一问题,提出一种鲁棒的图像匹配方法。通过Log-Gabor滤波器与图像卷积计算得到能够检测图像特征的相位一致性模型,接着考虑在各方向上的Log-Gabor卷积结果构造出具有旋转不变的标准化的最大索引图,最后在最大索引图上使用分布直方图完成特征描述符的构造以实现图像匹配。在具有辐射畸变的遥感图像上进行匹配实验,当待匹配图像发生旋转时,所提方法的正确匹配个数在50个以上,均方根误差(RMSE)为1.6~2.1像素,优于传统的辐射不变特征变换(RIFT)算法。实验结果表明,所提方法是一种有效的特征匹配算法。

Whether the image matching is good or not is an important factor affecting the quality of image stitching,and the traditional image matching method is sensitive to radiation distortion.In order to solve this problem,a robust image matching method was proposed.The phase congruency model which can be used to detect image features was obtained through the calculation of Log-Gabor filter and image convolution,and then the standardized maximum index map with rotation invariance was constructed by considering the results of Log-Gabor convolution in all directions.Finally,the distributed histogram method was used to construct feature descriptors on the maximum index map to achieve image matching.Matching experiments were carried out on remote sensing images with radiation distortion.When the image to be matched was rotated,the number of correct matches of the proposed method was more than 50,and the Root Mean Square Error(RMSE)was 1.6-2.1pixels,which is better than the traditional Radiation Invariant Feature Transformation(RIFT)algorithm.The experimental results shows that the proposed method is an effective feature matching algorithm.