摘要:
针对多波段SAR图像融合问题, 提出了一种基于contourlet变换的伪彩色图像融合方法. 该方法首先利用contourlet变换多尺度, 多方向以及各向异性的良好特性对不同波段SAR图像进行多尺度分解, 采用边缘信息量测因子策略融合方向高频子带, 加权平均策略融合低通子带. 然后对灰度融合图像利用混合高频提升滤波方法得到RGB彩色空间的三个颜色通道, 并在RGB空间量化显示,把人眼难以分辨的灰度信息转化为可分辨的色彩, 保持SAR图像的空间分辨率的同时增强了光谱分辨率. 采用Ku和L波段SAR图像进行融合实验, 并用客观评价因子对融合质量进行评价, 结果验证了该方法的有效性.
Abstract:
A false color image fusion method for multi-band synthetic aperture radar (SAR) images based on contourlet transform is presented. It works in the following way: firstly, the flexible multiresolution, high directionality and anisotropy of contourlet transform are used for implementing the multiscale decomposition for SAR images. In the contourlet transform domain, an edge information measurement rule is used to merge the directional subbands, and an averaging rule is used to merge the lowpass subbands. Then, a hybrid high boost filter algorithm is proposed to produce the red, green and blue color channels of red-green-blue (RGB) model based on the gray fused image. Finally, the false color fused image is displayed in the RGB color space. This method translates the gray information into color information available to human visual system and enhances the spectral resolution for SAR images. Experimental results is confirmed the validity of the proposed method.
下载:
