Space debris: pose estimation using stereo vision
| dc.contributor.advisor | Jordaan, H. W. | en_ZA |
| dc.contributor.advisor | Van Daalen, C. E. | en_ZA |
| dc.contributor.author | De Jongh, Wille Carel | en_ZA |
| dc.contributor.other | Stellenbosch University. Faculty of Engineering. Dept. of Electrical and Electronic Engineering. | en_ZA |
| dc.date.accessioned | 2019-02-22T06:24:25Z | |
| dc.date.accessioned | 2019-04-17T08:13:52Z | |
| dc.date.available | 2019-02-22T06:24:25Z | |
| dc.date.available | 2019-04-17T08:13:52Z | |
| dc.date.issued | 2019-04 | |
| dc.description | Thesis (MEng)--Stellenbosch University, 2019. | en_ZA |
| dc.description.abstract | ENGLISH ABSTRACT: Tracking the relative attitude and position of uncooperative in-orbit objects is vital for autonomous operations in space. Vision-based solutions have low power consumption and can provide a system with valuable information to perform pose determination. Estimation algorithms are required to extract the system states from visual measurements and many similar approaches have been investigated in mobile robotics. In this thesis, a chaser satellite is fitted with stereo cameras which are used to extract unique features on the surfaces of an uncooperative, unknown target. The scale invariant feature transform (SIFT) feature detector is used to identify and establish correspondence of the target features. A full state kinematic estimator is implemented using an extended Kalman filter (EKF) based on the simultaneous localisation and mapping (SLAM) approach. The filter makes use of the observations from the feature extractor to estimate the position and orientation of the target relative to the chaser along with the angular and linear velocities of the target. Shape and size reconstruction of the target is made possible using the sparsely tracked features. A simulation environment providing ground truth is used to verify the performance of the estimation algorithm. The integration of the estimator with the feature extractor is assessed using real world data. Experimental data is obtained from image sequences of a moving target in a laboratory set-up. Results show that the filter estimates the system states successfully and that the developed feature extractor is capable of detecting robust features with reliable correspondence. It is concluded that the use of a stereo-vision-based estimator is a viable option for autonomous operations such as space debris removal and satellite service missions. | en_ZA |
| dc.description.abstract | AFRIKAANSE OPSOMMING: Akkurate lokalisering van onbekende ruimte voorwerpe in verhouding tot ’n volgersatelliet is noodsaaklik vir outonome ruimte operasies. Skatting van die oriëntasie en posisie van die voorwerp deur middel van visuele sensors soos kameras, is ’n gewilde oplossing in die robotika-veld. Visuele sensors het ’n lae kragverbruik en is goedkoop om te implementeer. Lokaliseringsalgoritmes word benodig om die toestande van die voorwerp uit die visuele metings te onttrek. Hierdie tesis bespreek ’n stereo-kamera paar wat, saam met die skaal bestande kenmerk transform (SIFT) algoritme, gebruik word om unieke punte op die oppervlaktes van ’n nie-samewerkende, onbekende voorwerp te vind. Die algoritme is só ontwerp om rekord te hou van ooreenstemmende punte in opeenvolgende beelde. ’n Kinematiese toestands-skatter word geïmplementeer met behulp van ’n uitgebreide Kalman filter (EKF). Die skattingsalgoritme gebruik die gelyktydige lokalisering en kartering (SLAM) benadering. Die filter skat die relatiewe posisie en oriëntasie van die voorwerp af met betrekking tot die kameras. Die hoek- en lineêre-snelhede van die voorwerp word ook onttrek. ’n Verteenwoordiging van die voorwerp se grootte en vorm word saamgestel vanuit die geskatte posisies van die unieke voorwerp-punte. ’n Simulasie-omgewing, wat grondwaarheid voorsien, word gebruik om die werking van die skattingsalgoritme te toets. Die integrasie van die skatter met die beeldverwerkingsalgoritme word getoets deur gebruik te maak van eksperimentele beelde. Eksperimentele beelde word vasgelê deur ’n bewegende voorwerp waar te neem in ’n laboratorium-opstelling. Die stelsel toon bevredigende uitslae. Die EKF-SLAM benadering, in samewerking met die beeldverwerkingsalgoritme, is daartoe in staat om die voorwerp te lokaliseer relatief tot die kameras. Die studie kom tot die gevolgtrekking dat stereo-visiegebaseerde skatters voldoende is vir outonome ruimtesendings soos die diens van satelliete en die verwydering van ruimterommel. | af_ZA |
| dc.format.extent | 99 pages : illustrations | en_ZA |
| dc.identifier.uri | http://hdl.handle.net/10019.1/105812 | |
| dc.language.iso | en_ZA | en_ZA |
| dc.publisher | Stellenbosch : Stellenbosch University | en_ZA |
| dc.rights.holder | Stellenbosch University | en_ZA |
| dc.subject | Space Debris -- Removal of marine | en_ZA |
| dc.subject | UCTD | en_ZA |
| dc.subject | Space Debris | en_ZA |
| dc.subject | Kinematics | en_ZA |
| dc.subject | Autonomous agents (Computer software) | en_ZA |
| dc.title | Space debris: pose estimation using stereo vision | en_ZA |
| dc.type | Thesis | en_ZA |