基于独立分量分析的压缩域彩色图像融合

提出了一种基于独立分量分析的压缩域彩色图像融合方法,其基本思想是将源彩色图像进行IHS变换,将两幅图像的亮度分量/进行小波分解,对分解后的高频系数采用由独立分量分析得出的变换矩阵进行变换,然后对变换系数采用局部方差法进行融合,而对小波分解后的低频子图像分成若干N×N的图像块,分别对每个图像块做二维DCT变换,最后对融合的系数进行反变换得到融合结果,并用客观评价标准对结果进行了定量的分析.实验结果表明:该方法在提高空间信息的基础上,较少地降低了颜色失真.

A Color Image Fusion Approach Based on Independent Component Analysis in Compressed Domain

A color image fusion method based on compressed domain is proposed.First, IHS transformation is applied to the input images to obtain the I components.Second, low frequency coefficients after DWT are fused with DCT transformation, and the high frequency coefficients after DWT are applied with independent component analysis and then fused with area variance. Last, the fused coefficients are applied with inverse transformation to get the fused image.The experiment shows that this algorithm increases the information in the fused image while keeps the color distortion at lease.