9147

8465

كارشناسي ارشد

هوش مصنوعي و رباتيك

اصفهان: دانشگاه صنعتي اصفهان، دانشكده برق و كامپيوتر

1392

نه،100ص.: مصور،جدول،نمودار﴿رنگي﴾

ص.ع.به فارسي و انگليسي

مازيار پالهنگ

29/6/93

محمدرضا احمدزاده، عبدالرضا ميرزايي

مهندسي برق و كامپيوتر

ID8465

Scale Invariant Covariance Matrix Based Descriptor For 3d Point Cloud Registration Hossein Abedi Dorcheh hossein abedi@ec iut ac ir Supervisor Dr Maziar Palhang palhang@cc iut ac ir Date of submission 2014 01 13 Department of electrical and computer engineering Isfahan University of Technology Isfahan 84156 83111Iran Language Persian Degree M Sc Abstract3D modeling of the real world objects is one of the basic problems in the field of computer vision Not only this may be a goal in itself but also could be the prelude to solve other problems Theexisting methods typically employ depth sensors such as laser rang finders or stereo camerastoextract surface points of objects The set of extracted points is then called a point cloud Due tothenature of the sensors to extract the complete point cloud the sensor should be used atdifferentlocations and angles In this way multiple partial point clouds are obtained each of whichin adifferent coordinate system The problem of bringing all partial point clouds into acommoncoordinate system which is known as registration can be done by having at leastthreecorrespondences for each pair Since finding exact correspondences may not be possible manycorrespondences are computed at first and then RANSAC algorithm is used to find a nearoptimaltransformation To find the correspondences local 3d descriptors which are computed foreach point are used to describe the points around it By comparing alldescriptors of a pair of cloudswe are able to find the correspondences However the existingdescriptors have some problems suchas high computational and memory complexities and not beingscale invariant To tackle some ofthese problems a new covariance matrix based local descriptor isintroduced in this thesis Besideshaving well defined divergence measures with good theoreticalfeatures covariance matrices havesome other good features like high descriptivity non parametricityand low memory requirementwhich motivate us to use them as 3d descriptors In addition to introducing this new descriptor thisthesis also describes a new method for determining the points involved in thecomputation of aninstance of the introduced descriptor which makes it invariant to scale changes Experiments showthat the descriptor has a good descriptivity and really low computationalcomplexity Also it worthsnoting that the descriptor together with the neighbor determining method iscompletely invariant toscale changes a property which none of previous descriptors have

مازيار پالهنگ

محمدرضا احمدزاده، عبدالرضا ميرزايي