Skip to Main content Skip to Navigation

Tracking of hidden parts in video sequences and other problems raised by the 3D reconstruction of the urban environment

Romain Dupont
Abstract : This thesis deals with the urban environment reconstruction through the use of cameras and telemetric sensors. First, we focus on the spatiotemporal segmentation of video sequences in order to treat photographic data. We present a new technique of layer segmentation which extracts regions of similar parametric motion in a video sequence. It is based on temporal constraints defined and optimized over all images simultaneously and not successively, without any a priori on the observed scene. Taking advantage of temporal continuity, our framework considers both the visible and the hidden parts of each layer in order to increase robustness. The hidden parts of the layers are recovered, which could be of a great help in many high level vision tasks. Modeling the problem as a labeling task, we state it in a MRF-optimization framework and solve it with a graph-cut algorithm. Both synthetic and real video sequences show a visible layers extraction comparable to the one usually performed by state of the art methods, as well as a novel and successful segmentation of hidden layers. Secondly, we consider the use of heterogeneous telemetric and photographic data, in the same framework to obtain a 3D and textural reconstruction of the urban scenes. It has proved to be a powerful technique. A necessary condition to obtain good results is to calibrate accurately the single-row based telemetric sensor and the cameras together. We present a study of this calibration process and propose an improved extrinsic calibration technique. It is based on an existing technique which consists in scanning a planar pattern in several poses, giving a set of relative position and orientation constraints. The innovation is the use of a more appropriate laser beam distance between telemetric points and the planar target. Moreover, we use robust methods to manage outliers at several steps of the algorithm. Improved results on both theoretical and experimental data are given.
Document type :
Complete list of metadatas

Cited literature [153 references]  Display  Hide  Download
Contributor : Ecole Des Ponts Paristech <>
Submitted on : Tuesday, April 17, 2007 - 8:00:00 AM
Last modification on : Tuesday, April 17, 2007 - 8:00:00 AM
Long-term archiving on: : Thursday, September 30, 2010 - 8:06:54 PM


  • HAL Id : pastel-00002357, version 1



Romain Dupont. Tracking of hidden parts in video sequences and other problems raised by the 3D reconstruction of the urban environment. Mathematics [math]. Ecole des Ponts ParisTech, 2006. English. ⟨pastel-00002357⟩



Record views


Files downloads