Fusion of Panchromatic and Multispectral Images Based on the Second Generation Bandelet

In order to gain a new image with high spatial resolution and abundant spectrum by fusing panchromatic and multispectral images, a novel fusion algorithm based on the generation Bandelet is presented, and the fusion rule of Bandelet coefficients reposes on maximum absolute value of frequence. Fusion experiments based on new method, IHS and wavelet transform are carried out with panchromatic and multispectral images of Landsat-7. The experimental results show that fused image of new method is more excellent. Image edges are more distinct, and prove that Bandelet transform has the character of tracking image edges adaptively.

[1]  Stéphane Mallat,et al.  Discrete bandelets with geometric orthogonal filters , 2005, IEEE International Conference on Image Processing 2005.

[2]  Isabelle Bloch,et al.  Image Fusion , 1997 .

[3]  Jingwen Yan,et al.  A novel image fusion algorithm based on bandelet transform , 2007 .

[4]  Stéphane Mallat,et al.  Sparse geometric image representations with bandelets , 2005, IEEE Transactions on Image Processing.

[5]  M. Eismann Hyperspectral Remote Sensing , 2012 .

[6]  Chen Bo Image Coding Based on Second Generation Bandelet Transform , 2007 .

[7]  R. Sahoo,et al.  Hyperspectral Remote Sensing , 2013, Encyclopedia of Mathematical Geosciences.

[8]  D. Donoho Wedgelets: nearly minimax estimation of edges , 1999 .

[9]  G. Shaw,et al.  Signal processing for hyperspectral image exploitation , 2002, IEEE Signal Process. Mag..