Remote vessel detection using an optical imager and an AIS receiver combination for a CubeSat constellation
| dc.contributor.advisor | Steyn, H. W. | en_ZA |
| dc.contributor.author | Lombard, Francois | en_ZA |
| dc.contributor.other | Stellenbosch University. Faculty of Engineering. Dept. of Electrical and Electronic Engineering. | en_ZA |
| dc.date.accessioned | 2020-02-24T08:45:42Z | |
| dc.date.accessioned | 2020-04-28T12:10:41Z | |
| dc.date.available | 2020-02-24T08:45:42Z | |
| dc.date.available | 2020-04-28T12:10:41Z | |
| dc.date.issued | 2020-03 | |
| dc.description | Thesis (MEng)--Stellenbosch University, 2020. | en_ZA |
| dc.description.abstract | ENGLISH ABSTRACT: This thesis investigates illegal shing activities on the coast of South Africa. It was concluded that a constellation of 10 6U satellites orbiting in a specialised orbit, with repeating ground tracks, is best for surveying South Africa's coast. This orbit will require an orbital inclination angle of 48° at an altitude of 493:2 km. It can survey the coast with passive imagers up to 35 times a day, with a 10-day eclipse period in a solar year. Furthermore, it can send and receive radio signals in the South African region for 112 solar days with a maximum revisit time of 10 hours. Lastly, a comprehensive communication link analysis was conducted, and it is concluded that 95% of the vessels in the South African coast can be intercepted. Three orbits were compared using a simulation environment developed in MATLAB using a simplified general perturbation model. Additionally, the link analysis was performed by sampling the vessels from a realistic distribution and simulating the vessel and satellite antennas. It was found that 85% of the vessels lie on the satellite's radio horizon and it is recommended to have an antenna configuration with a stronger gain towards this region. Thereafter, classical image processing techniques, that can function on board a satellite, were investigated. Images were successfully pre-processed to extract regions of interests containing vessels. These regions were refined and classi ed using a linear support vector machine that extracts eight features from candidate regions. These were combined into six new features with less interclass variance, resulting in a classification score of 71%. To conclude, this thesis presents a successful feasibility study whether vessels can be detected in the South African coast using a 6U satellite constellation. It includes a comprehensive list of recommendations on which this thesis can be improved upon. | en_ZA |
| dc.description.abstract | AFRIKAANSE OPSOMMING: Hierdie tesis ondersoek onwettige vissery aktiwiteite aan die kus van Suid-Afrika. Daar is tot die gevolgtrekking gekom dat ’n konstellasie van 10 6U satelliete wat wentel in ’n gespesialiseerde wentelbaan, met herhalende grond spore, die suksesvolste is om die Suid-Afrikaanse kus te observeer. Hierdie wentelbaan benodig ’n hellingshoek van 48° met ’n hoogte bo seespieël van 493.2 km. Boonop sal hierdie wentelbaan die satellietkonstellasie in staat stel om die kus met ’n passiewe kamera 35 keer per dag te ondersoek met ’n 10-dag sonsverduistering-periode in een sonjaar. Verder, in die Suid-Afrikaanse omgewing, kan dit radioseine stuur en ontvang vir 112 son-dae met n herbesoek periode van 10 ure. Ten slotte, is ’n omvattende kommunikasiekanaal analise gedoen en daar is bevind dat 95% van die vaartuie in die Suid-Afrikaanse kus opgevang sal kan word. Drie wentelbane is vergelyk deur van ’n ontwikkelde MATLAB simulasie-omgewing gebruik te maak, wat ’n algemene vereenvoudigde versteuringsmodel implementeer. Verder is die kommunikasiemodel analise uitgevoer deur ’n realistiese steekproefruimte te skep wat gebaseer is op ’n realistiese verspreiding van vaartuie, asook om die vaartuie en satelliete se antennas te simuleer. Daar is bevind dat 85% van die vaartuie op die satelliet se radio-horison lê en dit word aanbeveel om ’n antenna-konfigurasie met ’n sterker aanwins gerig na die horison te gebruik. Daarna, is klassieke beeldverwerkingstegnieke ondersoek wat aanboord van ’n satelliet kan funksioneer. Beelde was suksesvol voorafverwerk om gebiede van belang, wat vaartuie bevat, te identifiseer. Hierdie gebiede se grense is verder verfyn en word geklassifiseer deur agt kenmerke uit die geïdentifiseerde gebiede te onttrek deur gebruik te maak van ’n linieêre masjienleer tegniek (support vector machine). Hierdie agt kenmerke word gekombineer om ses nuwes te skep met die doel om interklas variansie te verminder wat lei tot ’n klassifikasie telling van 72%. Ter samevatting, hierdie tesis bied ’n suksesvolle uitvoerbaarheidsstudie of vaartuie opgespoor kan word in die Suid-Afrikaanse kus deur gebruik te maak van ’n 6U satelliet konstellasie. Dit sluit ’n omvattende lys van aanbevelings in waarmee hierdie tesis se uitkomste verbeter kan word. | af_ZA |
| dc.description.version | Masters | en_ZA |
| dc.format.extent | xvii, 117, [20] leaves : illustrations (some color), maps (some color) | |
| dc.identifier.uri | http://hdl.handle.net/10019.1/107946 | |
| dc.language.iso | en | en_ZA |
| dc.publisher | Stellenbosch : Stellenbosch University | en_ZA |
| dc.rights.holder | Stellenbosch University | en_ZA |
| dc.subject | Ships -- Remote sensing | en_ZA |
| dc.subject | Ships -- Automatic identification systems | en_ZA |
| dc.subject | Satelite constellations | en_ZA |
| dc.subject | Coastal surveillance -- South Africa | en_ZA |
| dc.subject | Remote sensing | en_ZA |
| dc.subject | UCTD | en_ZA |
| dc.title | Remote vessel detection using an optical imager and an AIS receiver combination for a CubeSat constellation | en_ZA |
| dc.type | Thesis | en_ZA |