Projects:Student:Computer Vision:Stitching images of a flat surface

From FroboMind [www.frobomind.org]
Jump to: navigation, search

Project title

Stitching images of a flat surface from a moving unknown source.

Description

Many stitching algorithms are capable of determining the change in the camera position of two images that overlap (meaning finding the homogeneous transformation between camera positions) and thereby stitch images together. But what if you know that you are taking pictures of a flat (or close to flat) surface and want to rectify your images and simulate that you look straight down on the flat surface. The case for this problem are a moving robot taking images of a flat surface where several images cover the same part of the surface. The unknown factors are the camera position and the rectification matrix. The task is non trivial because you dont have calibrations images of chess boards or other objects with known parallel and correct positioned lines. Further the camera position may change over time.

Keywords

Computer vision, field robotics, weeding.

Level

MSc student project.

Project start

Anytime

Institution

Biosystems Engineering
Institute of Chemical Engineering, Biotechnology & Environmental Technology
University of Southern Denmark

Niels Bohrs Alle 1
DK-5230 Odense M
Denmark

If you don't live in Odense, we can provide contact to the dormitory administration as well as the nearby agricultural college, that has rooms for rent on a monthly basis.

Contact

Morten Laursen