Restoration of Multi-Spectral Images by Using the Panchromatic Image with Various Down-Sampling Methods

  • 朱 德和

Student thesis: Master's Thesis

Abstract

This thesis aims at the multi-spectral image error caused during the transmission of the remote sensing satellite image to combine the panchromatic image with multi-spectral B1 B2 B3 images as a way to generate predicted B ?_4 image for restoration In case the multi-spectral B1 B2 B3 images are used respectively to patch the B4 image though the structure of the band image is similar the difference of the reflective spectrum on four typical land objects (snow desert wetland vegetation) is significant that the image looked unnatural so the existing algorithm is applied to improve the current loss The panchromatic band covers most of the multi-spectral bands that the multi-spectral image data is included in the PAN image and through the fusion of the non-damaged multi-spectral images and the data of these images to generate the B ?_4 image prediction as a reference image to restore the multi-spectral image B4 The experimental results indicated that the algorithm may stably strengthen up the existing algorithm to better the image quality and allow the overall image look much natural also to present more details contained in the multi-spectral band image B4 clearly
Date of Award2015 Feb 12
Original languageEnglish
SupervisorShen-Chuan Tai (Supervisor)

Cite this

Restoration of Multi-Spectral Images by Using the Panchromatic Image with Various Down-Sampling Methods
德和, 朱. (Author). 2015 Feb 12

Student thesis: Master's Thesis