Remote Sensing Image Fusion Using Sparse Representation and Deep CNN

EasyChair Preprint 7842

Abstract

Remote sensing image fusion seems such as be an effective approach to make utilization of a big quantities of data through multiple sensors. Many remote sensing applications require both high spatial as well as high spectral resolutions, especially for GIS-based applications, but instead different earth satellites such as SPOT, Landsat 7, as well as IKONOS provide both panchromatic (Pan) images at a higher spatial resolution as well as multispectral (MS) images at a lower spatial resolution, and many remote sensing applications require both high spatial as well as high spectral resolutions. A remote sensing image fusion methodology is developed that seems to be utilize a deep convolutional neural network to absorb spectral and spatial features from source images. To generate a fused MS image having high spatial resolution, remote sensing image fusion can combine the spatial detail of a panchromatic (PAN) image with the spectrum information of a low-resolution multispectral (MS) image.

Keyphrases: CNN, Multispectral (MS) images., Panchromatic (Pan) images

@booklet{EasyChair:7842,
  author    = {Naluguru Udaya Kumar and Ayan Das Gupta and Syed Basha Shaik and Anooja B and Bhasker Pant and Jayaram Ramesh},
  title     = {Remote Sensing Image Fusion Using Sparse Representation and Deep CNN},
  howpublished = {EasyChair Preprint 7842},
  year      = {EasyChair, 2022}}