Related Experiment Video
Updated: Jul 31, 2026
08:25
Combining Eye-tracking Data with an Analysis of Video Content from Free-viewing a Video of a Walk in an Urban Park Environment
Published on: May 7, 2019
Vehicle detection by means of stereo vision-based obstacles features extraction and monocular pattern analysis
Gwenaëlle Toulminet1, Massimo Bertozzi, Stéphane Mousset
1Laboratoire Perception Systèmes Information, Université/INSA de Rouen, Mont-Saint-Aignan Cedex, France. gwenaelle.toulminet@insa-rouen.fr
Summary
This study introduces a stereo vision system for detecting preceding vehicles and calculating their distance. The system effectively identifies vehicles by analyzing 3-D features and matching them to a simplified vehicle model.
Area of Science:
- Computer Vision
- Robotics
- Automotive Engineering
Background:
- Accurate vehicle detection and distance estimation are crucial for advanced driver-assistance systems (ADAS).
- Traditional methods often struggle with varying environmental conditions and complex road scenes.
Purpose of the Study:
- To develop a robust stereo vision system for real-time detection and distance computation of preceding vehicles.
- To improve the accuracy and reliability of vehicle perception in automotive applications.
Main Methods:
- Utilized a stereo vision-based algorithm to extract three-dimensional (3-D) scene features.
- Implemented a feature selection process to isolate vertical objects, distinguishing them from the road and background.
- Employed a symmetry operator and a monocular vision-based approach for vehicle detection by matching features to a simplified vehicle model.
- Leveraged extracted 3-D information for precise distance computation.
Main Results:
- Successfully detected preceding vehicles using 3-D vertical features and model matching.
- Achieved accurate distance computation for detected vehicles based on stereo vision data.
- Demonstrated the system's capability to differentiate vehicles from background elements.
Conclusions:
- The proposed stereo vision system provides an effective solution for preceding vehicle detection and distance estimation.
- The method enhances the perception capabilities of ADAS by accurately identifying and localizing vehicles.
- This approach offers a foundation for more sophisticated autonomous driving systems.
