An Evaluation of an Omni- Directional (Fisheye) Vision Sensor for Estimating the Pose of an Object,

Two case studies of enhanced field of view are presented in this work. In the first an omni-directional lens is mounted on a single digital camera. In the second three perspective digital cameras are combined into a unique assembly (placed in row with known position to each another) to increase the sensing capabilities of the camera system. To generate precise pose information from a fisheye camera, the unified/ generic model is employed. The model is utilized to define the geometry of the camera. Omni-directional images captured are remapped to construct perspective images. A callback function for a mouse click event is developed to extract four non-coplanar feature points from remapped images of the target object. The image coordinates of the feature points are passed to a Pose from Orthography and Scaling with Iteration (POSIT) algorithm, to measure the objects pose in relation to the omni-directional system. The measured pose is exploited to calculate displacement and distance of the object from the camera. In this method, prior knowledge of the objects geometry is required. To calculate 3-d position from corresponding images, the 3-camra imaging system exploits Epipolar geometry and the Longuet-Higgins algorithm. In this technique in addition to individual camera calibration, position and orientation of each camera with respect to other cameras in the system are necessary. The experiments in this thesis are designed to measure the practicality of employing omni-directional lenses to generate metric information. In these experiments, measurements conducted via the 3-camera imaging system (control values) are compared against measurements gathered using the omni-directional imaging system.

Name thesis:degree_name

Master of Applied Science (MASc)

Level thesis:degree_level

Master's

Discipline thesis:degree_discipline

Engineering - Software Systems

Grantor dc:publisher

Faculty of Graduate Studies and Research, University of Regina

Year dc:date.issued

2015

Author dc:creator

• Parhizgar, Shahin

Advisor dc:contributor.advisor

• Paranjape, Raman

Committee members dc:contributor.committeemember

• Mehrandezh, Mehran
• Gelowitz, Craig
• Benedicenti, Luigi

Language dc:language.iso

en

DOI

https://doi.org/10.82465/4531

OAI identifier oai:identifier

oai:uregina.scholaris.ca:10294/5802