Toward RoboCup without Color Labeling

AI Magazine 

Unfortunately, color labeling can be applied to object-recognition tasks only in very restricted environments, where different kinds of objects have different colors. These stringent rules allow for simple mechanisms for object detection and recognition: Segment the captured image into blobs of the same color and interpret these blobs. To the best of our knowledge, all autonomous robot soccer teams with vision-based perception apply variations of such a method. However, because the RoboCup committee is planning to make the rules more realistic, these objectrecognition and -localization methods are becoming obsolete. The local statistics define an expectation of "how the two sides of the curve might look."

Similar Docs  Excel Report  more

TitleSimilaritySource
None found