论文部分内容阅读
Contourlet变换能有效地应用在图像增强领域,基于Contourlet变换的遥感影像融合算法常采用高频低频替换和频域线性加权两种融合方法,因此,通过粒子群优化算法对基于Contourlet变换分解不同子带图像所需的不同加权系数进行优化,采用多元回归分析方法设定目标函数,实现了全色影像与多光谱影像的融合,与传统的Contourlet变换、PCA算法、高通滤波遥感影像融合算法相比较,新方法在提高影像清晰度的同时在光谱保真度方面相对于其他算法有明显优势.
Contourlet transform can be effectively used in the field of image enhancement. The remote sensing image fusion algorithm based on Contourlet transform often uses two methods of high-frequency low-frequency replacement and frequency-domain linear weighting fusion. Therefore, by using the particle swarm optimization algorithm to decompose different sub-bands Different weighted coefficients needed for image optimization, the objective function was set by multiple regression analysis method, and the fusion of panchromatic image and multispectral image was realized. Compared with traditional Contourlet transform, PCA algorithm and high-pass filtering remote sensing image fusion algorithm, The new method has obvious advantages over other algorithms in terms of spectral fidelity while improving the image sharpness.