On-board, Fourier-Based Image-Analysis System for Satellite Observation of Gravity Waves

Abstract

The main objective of this thesis is obtaining data on gravity waves from images taken by the NTNU Test Satellite (NUTS). The satellite's main payload consists of a camera intended to capture images of gravity waves in the atmosphere. A limitation of the satellite is its ability to transmit data. Two separate approaches have been investigated to limit the amount of data that needs to be sent. One is the degree of image and video compression that can be applied to a series of images before data is corrupted. The second is implementing a 3D-Fast Fourier Transform (FFT) method that obtains the relevant data in the image series directly and only sends these. The NUTS-satellite is a project still in the making with a preliminary launch date in 2014. Currently there are no comparable, usable images of the wave phenomenon captured from space. Ground based images taken with fish-eye lenses are therefore used as a basis to make emulate satellite images. These are used to test how satellite images will be influenced by compression and how the 3D-FFT can extract the data. We found the 3D-FFT able to extract wave parameters from the images, providing significant reduction in the need for transmission of data compared to sending compressed images. In addition we found that the compression-approach corrupts data before a considerable reduction in transmitted data is achieved.