Zakaria, Mohd Maarof Mohd
(2016)
A Vision Guided Method Using Omnidirectional Camera.
Masters thesis, Universiti Sains Malaysia.
Abstract
The omnidirectional camera is widely used in robotics technology to get the image of the overall direction and many studies using a 360 ° view. In this project, a new guided localization method using the omnidirectional camera is developed. The task is using two omnidirectional cameras with 360 vision to capture images upside and down or similar as a global vision camera that can capture images in 720. By using the omnidirectional camera and local cues, a robotic platform will be able to determine its current position and utilise it for onward navigation. An image-based algorithm is developed to overcome the problem of scaling and distortion of the object within the surrounded image, and use it to determine its exact location referring to a pre-determined reference marker. All the system will try to reduce distortion of images that have been captured by this two omnidirectional cameras. This project will involve a development of hardware and software integration, with emphasises to be given on the solving the localization problem. The result from this studiey is the images from two cameras can be processed to form 720° image and coloured object detection can be done.
Actions (login required)
 |
View Item |
![[img]](http://eprints.usm.my/style/images/fileicons/application_pdf.png)
