High Accuracy Volume Measurement of Small Shapes using Stereo Vision
Independent thesis Advanced level (professional degree), 20 credits / 30 HE creditsStudent thesis
This thesis addresses the possibility of determining the exact volume of a small irregularshape using imaging methods. More specifically the suggested method is using a photogrammetric technique called Stereo Vision which makes it possible to extract 3-dimensional information from two or more digital images displaying different viewsof the same object.The motivation is to develop a non-destructive way for the Swedish mining companyLKAB to measure the porosity of their produced iron ore pellets. With known mass and absolute density the volume uniquely defines the porosity of a sample. With high accuracy volume measurements one can guarantee a certain quality of the porosity estimations.Techniques exist today that measure the porosity of individual pellets but they are either classified as environmentally unfriendly or affect the physical propertiesof the measured sample.The designed system is using the computer vision library OpenCV together with the mesh manipulation software MeshLab to generate an high accuracy model of the photographedobject. The volume of this model is then estimated using volume integrals according to already developed algorithms.Testing indicates that the final system can determine the volume with a relative error of approximately 1-2 %. As the targeted error is below 0.28 % this means that the system as it is designed today produces less accurate results than other already existing methods for porosity measurements. Future improvements are however suggested how to further increase the accuracy of this method to potentially outperformits competitors.
Place, publisher, year, edition, pages
2012. , 90 p.
Technology, Stereo Vision, pellets, volume, LKAB, photogrammetry, OpenCV, accuracy, non-destructive
Teknik, Stereo Vision, Pellets, Volume, LKAB, Photogrammetry, OpenCV, Accuracy, Non-destructive
IdentifiersURN: urn:nbn:se:ltu:diva-56322Local ID: d1aa8c3c-219b-4110-841f-6a332b80e3b4OAI: oai:DiVA.org:ltu-56322DiVA: diva2:1029709
Subject / course
Student thesis, at least 30 credits
Computer Science and Engineering, master's level
Validerat; 20120821 (anonymous)2016-10-042016-10-04Bibliographically approved