Abstract
In RoboCup, all objects in the game are color coded to
allow easy segmentation and identification. However, color vision is very sensitive to the exact color of the illumination which makes the identification task much harder than is often anticipated. Algorithms that work
in the lab in fluorescent light may not function at all in an actual game where incandescent lights are used.
This article is intended as a short tutorial on color vision and demonstrates a method for retrieving the color of image patches independently of the color of the illumination. The computational scheme is tested on a set of images of orange balls on a green background with varying illumination.
allow easy segmentation and identification. However, color vision is very sensitive to the exact color of the illumination which makes the identification task much harder than is often anticipated. Algorithms that work
in the lab in fluorescent light may not function at all in an actual game where incandescent lights are used.
This article is intended as a short tutorial on color vision and demonstrates a method for retrieving the color of image patches independently of the color of the illumination. The computational scheme is tested on a set of images of orange balls on a green background with varying illumination.
| Original language | English |
|---|---|
| Publisher | Lund University Cognitive Science |
| Number of pages | 7 |
| Volume | 98 |
| Publication status | Published - 2003 |
Publication series
| Name | LUCS Studies |
|---|---|
| Volume | 98 |
| ISSN (Print) | 1101-8453 |
Bibliographical note
The information about affiliations in this record was updated in December 2015.The record was previously connected to the following departments: Cognitive Science (015001004), Neuronano Research Center (NRC) (013210020), Electrical and information technology (011041010), Zoology (Closed 2011) (011012000)
Subject classification (UKÄ)
- Computer Vision and Robotics (Autonomous Systems)