Related Experiment Video
Updated: Apr 24, 2026
10:04
Sample Drift Correction Following 4D Confocal Time-lapse Imaging
Published on: April 12, 2014
15.7K
Sequential and automatic image-sequence registration of road areas monitored from a hovering helicopter
Fatemeh Karimi Nejadasl1, Roderik Lindenbergh2
1Fugro Intersite B.V., Fugro, Dillenburgsingel 69, 2263 HW Leidschendam, The Netherlands. fkariminejadasl@gmail.com.
Sensors (Basel, Switzerland)
|September 9, 2014
Summary
This study introduces an automatic image registration method for road surveillance, crucial for vehicle tracking. The technique accurately aligns sequences, even with motion, achieving zero pixel error for enhanced traffic analysis.
Area of Science:
- Computer Vision
- Image Processing
- Robotics
Background:
- Automated vehicle tracking requires precise image registration from aerial surveillance.
- Traffic and weather conditions necessitate robust methods for handling scene-induced motion.
Purpose of the Study:
- To develop an automatic and sequential image registration method for road surveillance.
- To enable accurate vehicle tracking by addressing scene-induced motion in image sequences.
Main Methods:
- A novel framework for automatic and sequential image registration was designed and implemented.
- Registration errors are sequentially corrected using two homography assumptions.
- An approximate registration is refined using a restricted search area for efficiency.
Main Results:
- The method was tested on a 1500-image sequence, demonstrating its effectiveness.
- Achieved registration between arbitrary images with zero geometric error in pixel accuracy.
- Successfully handled scene-induced motion, a common challenge in road surveillance.
Conclusions:
- The proposed method provides accurate and automated image registration for road surveillance.
- This technique is a valuable component for advanced vehicle tracking systems.
- The zero-pixel error accuracy highlights the method's robustness and reliability.
