A model-based tracking framework for textureless 3D rigid curved objects

Abstract

This paper addresses the problem of tracking texturelessrigid curved objects. A common approach uses polygonalmeshes to represent curved objects inside an edge-based trackingsystem. However, in order to accurately recover their shape,high quality meshes are required, creating a trade-off betweencomputational efficiency and tracking accuracy. To solve thisissue, we suggest the use of quadrics calculated for each patchin the mesh to give local approximations of the object contour.This representation reduces considerably the level of detail of thepolygonal mesh while maintaining tracking accuracy. The noveltyof our research lies in using curves to represent the quadrics’projection in the current viewpoint for distance evaluation insteadof comparing directly the edges from the mesh and detectededges in the video image. In our tracking framework, we alsoinclude a method to calculate the measurable Degrees of Freedom(DoF) of the target object. This is used to recover the poseparameters when the object has less than 6DoF. Experimentalresults compare our approach to the traditional method ofusing sparse and dense meshes. Finally, we present a potentialAugmented Reality application of the proposed method.