3-D motion estimation, understanding, and prediction from nosiy image sequences
IEEE Transactions on Pattern Analysis and Machine Intelligence
A Four-step Camera Calibration Procedure with Implicit Image Correction
CVPR '97 Proceedings of the 1997 Conference on Computer Vision and Pattern Recognition (CVPR '97)
| Hi-index | 0.00 |
Conventional identification needs system input and output recorded by sensors. In some situations, installation of a conventional sensor may be difficult or not possible due to hazardous environment and/or confined area. This paper proposes a new approach to system identification via image processing techniques. This approach permits non-intrusive and remote identification. One or more cameras can be an alternative to conventional sensors. Dynamical information of the system can be extracted from the recorded images. Details of information extraction from images are presented in this paper. For linear models, conventional identification techniques based on regression analysis are applied. For nonlinear models, the adaptive tabu search (ATS), one of the AI search techniques, is employed. The proposed approaches have been tested against the cart-plus-pendulum (CPP) system, and the vibrating tube system (VTS), respectively. Practical results have been achieved with high satisfaction.