Navigation and Image Processing: Object Tracking in Image Sequences
Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis
This report presents the work conducted at Sagem (Safran) in the context of my master thesis internship. In the navigation systems development centre, we try to replace the GPS-hybridisation of inertial systems with a vision-hybridisation in order to remove any satellite dependence and get more autonomous. The project presented here deals with the problems of extrinsic calibration of the camera. The chosen solution consists in tracking points of known asolute position (window corners in our work) in a video sequence and using them to make the harmonizaton with the inertial system as accurate as possible. This report contains the computer vision work implemented during the thesis: a bibliography focused on robust features extraction and tracking is detailed an explains how we came out with the technique for tracking the window corners. This method exploits the strengths of Harris´corner detector to detect the corners, and the invariance of the SIFT points to locate them in the image and thus be able to track them. This algorithm shows a good performance for our problem and fully meets the needs of the extrinsic calibration.
Place, publisher, year, edition, pages
EES Examensarbete / Master Thesis, XR-EE-KT 2013:001
Electrical Engineering, Electronic Engineering, Information Engineering
IdentifiersURN: urn:nbn:se:kth:diva-141416OAI: oai:DiVA.org:kth-141416DiVA: diva2:696657