Pansharpening of Satellite Images with Convolutional Sparse Coding and Adaptive PCNN-Based Approach

Remote Sensing Pansharpening with TV-H−1 Decomposition and PSO-Based Adaptive Weighting Method

“…Different methods as stated are applied to different datasets, and resultant images are shown in succeeding figures in which (a-l) is as follows,(a) image is degraded PAN image, (b) is Reference MS image ,(c) resultant image after applying Exponential method, (d) resultant image of applying band dependent spatial details injection method ,(e) resultant image of Brovey transform with Haze correction method,(f) resultant image for gram-Schmidt (GS) transformation ,(g) fused image for NSCT+SR transformation [24] ,(h) is fused image for pansharpening by Convolutional Neural Networks(PNN) [37] ,(i) resultant image for target-Adaptive CNN-Based Pansharpening (PNNplus) [38] ,(j) is resultant image after applying total variation method [31],(k) is resultant image for CSC+APCNN method [48] .Methods used for comparison (c,d,e,f,h,i,i,j) directly taken from package [45] At reduced scale (wald's protocol synthesis property): At reduced scale four different datasets are used. As per the synthesis property [46] , I.…”

“…The proposed method pansharpening with TV-H -1 decomposition and adaptive weighting method is compared with existing pansharpening methods. Which includes EXP method where image is interpolated with 23 coefficients for fusion ,Band dependent spatial details injection (BDSD) method , Brovey transform with haze correction (BT-H) , Gram-Schmidt based approach (GS) , Non sub-sampled contourlet transform with sparse representation based fusion (NSCT+SR) ,Pansharpening with CNN (PNN) , Pansharpening with target adaptive CNN (PNNplus),Total variation (TV),Convolutional sparse coding with pulse adaptive PCNN (CSC+ APCNN) [48].…”