Please use this identifier to cite or link to this item:
Title: Improvement of micro-satellite multispectral pushbroom sensor band co-registration: An XSAT case study
Authors: Tan, W.J. 
Kwoh, L.K. 
Keywords: Interband co-registration
Micro Satellite
Platform stability
Issue Date: 2012
Citation: Tan, W.J.,Kwoh, L.K. (2012). Improvement of micro-satellite multispectral pushbroom sensor band co-registration: An XSAT case study. 33rd Asian Conference on Remote Sensing 2012, ACRS 2012 2 : 1593-1599. ScholarBank@NUS Repository.
Abstract: Small sized, lightweight micro satellites are increasingly common within the earth observation community. This class of satellite can offer a good balance of payload capacity and low deployment cost. XSAT is a micro satellite designed and built by Nanyang Technological University, Singapore which carries a multispectral optical pushbroom sensor with NIR, red and green channels and a designed GSD of 10m. A known and expected tradeoff with smaller satellites is a lower level of platform stability. If the payload is a pushbroom multispectral camera, this instability would result in band co-registration challenges if a rigid camera model is the only model used to perform band stacking. This problem is compounded if high accuracy, high timing resolution pointing information is not known and recorded at imaging time. Micro satellite attitude instability is primarily a combination of lower frequency attitude control loop effects, and higher frequency jitter due to mechanical vibration. To improve image band stacking, XSAT's image processing software includes two components: 1) a variable order polynomial model to remove the lower frequency drifting, and 2) an inter-band feature matching algorithm to remove the higher frequency jitter. The first component has been described in a separate paper published in ACRS2011 and ISPRS2012. In this paper we describe work done on the second component. This component implements a feature matching algorithm in moving search windows placed at evenly spaced intervals throughout the image. Within these windows, feature edges are extracted using a Sobel edge detection operator and matching features are located in other bands by searching for locations in them with the highest cross correlation coefficient between the Sobel operator outputs for the two bands. Finally, a table of offsets is built, and interpolation is performed to produce a resampled band stacked image with the jitter accounted for. Results for this experimental approach are presented.
Source Title: 33rd Asian Conference on Remote Sensing 2012, ACRS 2012
ISBN: 9781622769742
Appears in Collections:Staff Publications

Show full item record
Files in This Item:
There are no files associated with this item.

Page view(s)

checked on Dec 2, 2021

Google ScholarTM



Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.