基于小波包移频算法的遥感图像融合技术

为了对分类最小二乘支持向量机实施有效的稀疏化，以提高分类速率，采用分类相关分析算法，按序提取样本核矩阵的全部分类相关成分，并依据样本核矩阵各列与分类相关成分的相关性，对训练集所有个体按分类的重要性排序，进而可选取最重要的部分个体作为支持向量，并将其余非支持向量的信息转移至支持向量，以提高支持向量的分类表达能力.由此构建一种新的稀疏型最小二乘支持向量机CS LSSVM，并将其应用于多个模式分类的实际问题.测试结果表明，CS LSSVM稀疏性很强，且保持了标准LSSVM的分类性能，还可直接适用于多类问题.

Remote sensing image fusion method based on wavelet packet frequency-shift

A wavelet packet frequency-shift algorithm for remote sensing image fusion was introduced to solve the problem of serious frequency aliasing caused by dyadic-down sampling.When images were decomposed in the wavelet packet space,the maximal-absolute-value method was employed to calculate high-frequency coefficients, and low-frequency coefficients were calculated by the local energy method.Entropy, cross entropy,spectrum distorting index and edge index were used to estimate the quality of the result image generated by fusing the enhanced thematic mapper (ETM ) multi-spectrum image and the ETM panchromatic image.Experimental results showed that the proposed method is better than common wavelet packet fusion methods in preserving spectral characteristics and retaining high spatial resolution.