Sub-pixel image translation estimation on a nanosatellite platform

Jurgen, Ludemann (2019-04)

Thesis (MEng)--Stellenbosch University, 2019.

Thesis

ENGLISH ABSTRACT: Nanosatellites are limited in their physical size, which limits the physical size of payloads they can carry, thereby limiting the quality of images taken during CubeSat Earth Observation missions. Algorithms exist that combine partially overlapping images to produce better output image quality. These algorithms may either improve the signal-to-noise ratio via averaging, increase resolution via super-resolution or merely remove redundant information via mosaicing. Typically, they only function properly if the geometric transformations between consecutive images are known with high accuracy. They can either be applied terrestrially or on-board a satellite. Downloading large raw image data sets for terrestrial processing is impractical for a CubeSat mission, and therefore an on-board solution is desirable. This thesis discusses the accurate determination of the transformation between consecutive images on-board, laying the foundation for e cient onboard de-noising, super-resolution and mosaicing. Two common methods used to determine translation { normalised cross correlation (NCC) and phase correlation { are investigated. From simulated results, NCC is shown to be the better candidate for our application. NCC achieves sub-pixel accuracy by making use of polynomial least squares regression. NCC is well suited for implementation on a satellite platform where images are captured in quick succession, resulting in partially overlapping images with little rotation between frames. We compare two potential hardware platforms { the MicroZed 7020 and Jetson TK1 { and then describe how we implemented our proposed solution onto the former, using a hardware description language. Software simulation and rmware-implementation results, using simulated data, are compared and discussed. Subsequently, the MicroZed 7020's implemented design is characterised, compared and discussed in terms of algorithm and platform performance.

AFRIKAANSE OPSOMMING: Nanosatelliete is beperk t.o.v hul grootte, gevolglik is die `loonvrag' wat hulle kan dra ook beperk. Dit het 'n negatiewe uitwerking op die kwaliteit van die beelde, wat tydens waarnemings missies van CubeSats gemaak word. Algoritmes bestaan wat gedeeltelik oorvleuelende beelde kombineer om vir hoer kwaliteit `uitset' beelde te sorg. Sulke algoritmes kan die sein-tot-ruis verhouding verbeter via `beeld-sommering', die resolusie verhoog deur super-resolusie of oorbodige informasie via `mosaïekmetodes' verminder. Sodanige algoritmes funksioneer slegs optimaal wanneer die geometriese transformasies tussen agtereenvolgende beelde tot 'n hoë vlak van akkuraatheid bepaal word. Die soort algoritmes kan aan boord, of op die grond toegepas word. Dit is onprakties is om sulke groot ongeformatteerde datastelle af te laai vir prossessering op die grond tydens die missie van 'n Nanosatelliet, dus geniet 'n aanboord oplossing voorkeur. Hierdie tesis bespreek die akkurate bepaling van inter-beeld transformasies aan bord boord van 'n satelliet. Dit l^e die fondament vir aanbord sein-suiwering, super-resolusie en mosaïek metodes. Twee algemene metodes { genormaliseerde kruiskorrelasie (GK) en fasekorrelasie { word ondersoek. Simulasieresultate dui aan dat GK meer doeltreffend is vir ons doeleindes. GK behaal sub-pixel akkuraatheid deur middel van polinomiese kleinste kwadraat regressie. GK is geskik vir 'n platform waar opeenvolgende deels-oorvleuelende beelde, met weglaatbaar klein relatiewe rotasie, intyds bewerk moet word. Ons vergelyk twee potensieële hardeware platforms { die Jetson TK1 en die MicroZed 7020 { en implementeer die voorgestelde oplossing op die laasgenoemde, met die gebruik van hardewarebeskrywingskode. Die resultate van die sagtewaresimulasie word met die geimplementeerde hardewareresultate vergelyk, met die gebruik van gesimuleerde data. Die MicroZed 7020 se geïmplementeerde ontwerp word ook uiteengesit, vergelyk en bespreek, in terme van die vermoë van die algoritme en hardewareplatform.

Please refer to this item in SUNScholar by using the following persistent URL: http://hdl.handle.net/10019.1/105831
This item appears in the following collections: