多光谱遥感影像亮度空间相位一致性特征点检测

Interest Point Detection for Multispectral Remote Sensing Image Using Phase Congruency in Illumination Space

提出了一种基于亮度空间和相位一致性理论的多光谱遥感影像特征点检测算法。首先利用参数自适应的灰度变换函数建立影像亮度空间;然后结合相位一致性方法在影像亮度空间进行候选特征点检测,并将候选特征点映射到原始影像上进行非极大值抑制;最后在尺度空间计算特征点的特征尺度值。本文方法有效结合了亮度空间特征检测和相位一致性特征检测的优势,对多光谱遥感影像的辐射变化具有较强的稳健性。试验结果证明,与传统特征点检测算法相比,本文方法在特征重复率和重复特征数量方面都具有明显的优势。

A robust interest point detection algorithm based on illumination space and phase congruency is proposed in this paper.Firstly,image illumination space is constructed by using aparameters adaptive method.Secondly,aphase congruency based interest point detection algorithm is adopted to compute candidate points in illumination space.Then,all interest point candidates are mapped back to the original image and a non-maximum suppression step is added to find final interest points.Finally,the feature scale values of all interest points are calculated based on the Laplacian function.The proposed algorithm combines the advantages of illumination space and phase congruency,which makes the proposed method robust to the radiation variation of multispectral images.The experimental results show that the proposed method performs better than other traditional methods in feature repeatability rate and repeated features number.