Paper
17 February 2012 A new 3D-registration method based on arbitrary geometric shape and unscented particle filter
Min Zhao, Bingqi Liu, Dongsheng Wu
Author Affiliations +
Abstract
A tracking algorithm based on arbitrary geometric shape and Unscented Particle Filter is proposed in this paper. Traditional tracking algorithm based on arbitrary geometric shape has following advantages, matching convenient, high precision, but when the geometric shape is not fully visible it results tracking failure so that restricts its use. Another tracking algorithm based on natural feature utilizes natural feature points to finish registration, there are disadvantages such as error accumulation as time goes by. This approach firstly uses arbitrary geometric shape to finish camera calibration, then use unscented particle filter to predict the possible location of the camera, and get 3-D registration matrix, it can reduce non-linear, non-Gaussian interfere brought by other issues. Experiment results show that this method can efficiently work under the circumstances of occlusion, large changes in illumination, and have a high accuracy.
© (2012) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Min Zhao, Bingqi Liu, and Dongsheng Wu "A new 3D-registration method based on arbitrary geometric shape and unscented particle filter", Proc. SPIE 8332, Photonics and Optoelectronics Meetings (POEM) 2011: Optoelectronic Sensing and Imaging, 83320Y (17 February 2012); https://doi.org/10.1117/12.918800
Advertisement
Advertisement
RIGHTS & PERMISSIONS
Get copyright permission  Get copyright permission on Copyright Marketplace
KEYWORDS
Cameras

Particle filters

Detection and tracking algorithms

Calibration

Imaging systems

Computing systems

Optical filters

Back to Top