The Förstner Interest Point Operator Subwindow Localization on SIFT Keypoints
Independent thesis Advanced level (degree of Master (Two Years)), 300 HE creditsStudent thesis
This thesis suggests a modiﬁcation to the popular Scale Invariant Feature Transform (SIFT) algorithm (Lowe, 2004) often used in photogrammetry and computer vision to ﬁnd features in images for measurement. The SIFT algorithm works by ﬁrst detecting points in images at diﬀerent scales and sizes. It then reﬁnes the position of the found points. The algorithm creates a descriptor of the point based on the region around the point. Finally the points can be matched against other points in diﬀerent images using the descriptor. The suggested modiﬁcation is built upon a paper by Förstner and Gülch (1987) where a method for performing a subwindow localization is presented. In this thesis the keypoints detected by the SIFT algorithm are modiﬁed on the subwindow level in order to improve the robustness with respect to the selected window position. Several diﬀerent methods of tweaking the suggested modiﬁcation and the SIFT algorithm were tested. The methods were evaluated on two diﬀerent test cases. The ﬁrst used a camera calibration software to compare accuracy of keypoints by looking at the residuals of the calibration. The other test involved creating a point cloud of images of a planar surface, evaluating the results by looking at the standard deviation in keypointoﬀset from the plane.The results show that neither test gave evidence that the proposed modiﬁcation was an improvement. It was found that the algorithm had problems with oblique projections of circles, i.e. ellipses. Therefore there is potentialto use homography in special cases to circumvent this problem and get better precision. Furthermore tests involving more lines and intersections in the test images should be performed before this suggested modiﬁcationcan be completely discarded.
Place, publisher, year, edition, pages
2015. , 28 p.
Photogrammetry, Subwindow, Localization, SIFT
IdentifiersURN: urn:nbn:se:umu:diva-105492OAI: oai:DiVA.org:umu-105492DiVA: diva2:825802
Subject / course
Examensarbete i teknisk fysik
Master of Science Programme in Engineering Physics
2015-06-12, NA332, Umeå, 14:00 (English)
Börlin, Niclas, Universitetslektor
Björklund, Johanna, UniversitetslektorServin, Martin, Universitetslektor