|
Title:
|
PRECISE 3D MEASUREMENTS FOR TRACKED OBJECTS FROM PAIRS OF SYNCHRONIZED VIDEO SEQUENCES |
|
Author(s):
|
Panagiotis Agrafiotis, Andreas Georgopoulos, Anastasios D. Doulamis, Nikolaos D. Doulamis |
|
ISBN:
|
978-989-8533-25-8 |
|
Editors:
|
Hans Weghorn |
|
Year:
|
2014 |
|
Edition:
|
Single |
|
Keywords:
|
3D measurements, tracking, stereovision, bundle adjustment, camera calibration |
|
Type:
|
Full Paper |
|
First Page:
|
101 |
|
Last Page:
|
108 |
|
Language:
|
English |
|
Cover:
|
|
|
Full Contents:
|
click to dowload 
|
|
Paper Abstract:
|
The purpose of this work is to design a system suitable to perform precise and fast 3D measurements from a pair of synchronized video sequences obtained from a stereo configuration. To this direction we combine a robust tracker with photogrammetric techniques into a fast and reliable system. For tracking objects and people, we introduce a stable human tracker able to cope efficiently with the trade-off between model stability and adaptability. In more detail, we adopt probabilistic mixture models like the Gaussian Mixture Models (GMMs) which exploit geometric properties for background modelling. Then, we integrate iterative motion information methods, concerned by shape and time properties, to estimate image regions of high confidence for updating the background model. In order to achieve accurate and precise 3D measurements, a camera calibration procedure was applied in order to recover the parameters of the intrinsic orientation of the cameras of the stereo configuration. Each camera calibration was performed using a 2D control field being moved by a person and by using template matching techniques. Finally, for calculating the 3D trajectory of the moving person with precision and reliability, we apply bundle adjustment for the frames from the left and right camera. Bundle adjustment that is generally used as a final step of any reconstruction algorithm is very accurate and has the advantages of being tolerant of missing data while providing a true Maximum Likelihood estimate. The results have been ested and evaluated in real life conditions for proving the robustness and the accuracy of the system. |
|
|
|
|
|
|
