Hier finden Sie wissenschaftliche Publikationen aus den Fraunhofer-Instituten.

Camera Calibration for Color Classification

: Urbann, Oliver; Menges, Dino; Schwarz, Ingmar; Tasse, Stefan; Stenzel, Jonas


Institute of Electrical and Electronics Engineers -IEEE-:
4th Asia-Pacific Conference on Intelligent Robot Systems, ACIRS 2019 : 13-15 July 2019, Nagoya, Japan
Piscataway, NJ: IEEE, 2019
ISBN: 978-1-7281-2229-8
ISBN: 978-1-7281-2228-1
ISBN: 978-1-7281-2230-4
Asia-Pacific Conference on Intelligent Robot Systems (ACIRS) <4, 2019, Nagoya>
Fraunhofer IML ()
ACRIS; machine learning; image processing; evolutionary strategies

This paper is about the optimization of camera parameters to achieve good color classification for a limited amount of colors. In the context of RoboCup it is very important that the robots recognize the field, the ball and the competitors. This recognition can be done using the objects' colors. All approaches to color classification are dependent on the quality of the provided camera picture. A manual optimization of camera parameters is time consuming and we noticed that optimal parameters can be very different from manually chosen ones. Therefore we developed a general automatic approach to optimize camera parameters for color classification, based on evolutionary algorithms.