基于光谱空间变换的遥感图像目标探测方法研究

针对遥感图像在频率域中的表征，提出了一种基于光谱空间变换的遥感图像目标探测方法。该方法首先利用傅里叶变换，将遥感图像从空域转变到频率域;然后利用频谱能量楔状采样和谐波叠置等手段，将不同频谱能量所表征的目标特征信息分解到不同的高、低频段中，由此获取对应目标特征在频率域中的探测标志;最后结合在频谱能量上具有方向和频带选择性的匹配Gabor滤波器，实现了居民楼地物目标的有效探测。试验结果表明，文章所提出的方法能够较好地探测遥感图像的目标信息，并且具有特定方向上目标检测的能力。

A Method of Object Detection for Remote Sensing Imagery Based on Spectral Space Transformation

Object detection is an intermediate link for remote sensing image processing, which is an important guarantee of remote sensing application and services aspects. In view of the characteristics of remotely sensed imagery in frequency domain, a novel object detection algorithm based on spectral space transformation was proposed in the present paper. Firstly, the Fourier transformation method was applied to transform the image in spatial domain into frequency domain. Secondly, the wedge-shaped sample and overlay analysis methods for frequency energy were used to decompose signal into different frequency spectrum zones, and the center frequency values of object’s features were acquired as detection marks in frequency domain. Finally, object information was detected with the matched Gabor filters which have direction and frequency selectivity. The results indicate that the proposed algorithm here performs better and it has good detection capability in specific direction as well.