Vehicle Class Recognition with Probes Using 3D Curves

Han, Dongjin

Full Metadata

Overview

Year:: 2009
Contributor:: Han, Dongjin (creator); Cooper, David (director); Kimia, Benjamin (reader); Mundy, Joseph (reader); Brown University. Division of Engineering. Electrical Sciences and Computer Engineering (sponsor)
Genre:: theses
Subject:: Class; Recognition; Probes; 3D; Structure from Motion; Vehicles
Extent:: xii, 101 p.
DOI: https://doi.org/10.7301/Z01J980M

Files

Description

Abstract:: We present new methods of vehicle class recognition for four classes of vehicles (specifically, SUV, mini-van, sedan, and pickup truck) using one or more fixed video-cameras in arbitrary positions with respect to a road. The ultimate goal is to deal with a very large number of classes. The road is assumed to be essentially straight. The system works as follows: 1. A vehicle silhouette is computed, using standard algorithms, in each video frame. 2. The vehicle straight line trajectory and the 3D position along the trajectory, as seen in each of a sequence of video frames, is estimated using a new computationally simple approach based on vanishing points and the cross ratio invariance. 3. The 3D points are computed from the vehicle apparent contours in a sequnece of frames. 4. 3D geometry such as total length, cabin length, width, height, and functions of these are computed and become features for use in a classifier. 5. Classification is done by a minimum probability of error recognizer. 6. The system is designed to produce good classification even when portions of the silhouette in an image frame may not be good by extensive use of histograms. 7. Finally, since classification error based on a single video clip may not be small enough, where additional video clips taken elsewhere are available we design classifiers based on two or more video clips, and this results in significant classification-error reduction.
Notes:: Thesis (Ph.D.) -- Brown University (2009)

Content

Access Conditions

Rights: In Copyright
Restrictions on Use: Collection is open for research.

Citation

Han, Dongjin, "Vehicle Class Recognition with Probes Using 3D Curves" (2009). Engineering Theses and Dissertations, Electrical Sciences and Computer Engineering Theses and Dissertations. Brown Digital Repository. Brown University Library. https://doi.org/10.7301/Z01J980M

Relations

Collections:

Engineering Theses and Dissertations

Theses and Dissertations for the Engineering department.
...
Electrical Sciences and Computer Engineering Theses and Dissertations

Theses and Dissertations for the Electrical Sciences and Computer Engineering department.
...