Object Recognition from Local Scale-Invariant Features
ICCV '99 Proceedings of the International Conference on Computer Vision-Volume 2 - Volume 2
Computer Graphics and Geometric Modelling: Implementation & Algorithms
Computer Graphics and Geometric Modelling: Implementation & Algorithms
Robust Object Recognition with Cortex-Like Mechanisms
IEEE Transactions on Pattern Analysis and Machine Intelligence
Information-based compact pose SLAM
IEEE Transactions on Robotics
Fusing Monocular Information in Multicamera SLAM
IEEE Transactions on Robotics
Active vision for sociable robots
IEEE Transactions on Systems, Man, and Cybernetics, Part A: Systems and Humans
Hi-index | 0.00 |
This paper proposes a fast colour-based object recognition and localization for soccer robots. The traditional HSL colour model is modified for better colour segmentation and edge detection in a colour coded environment. The object recognition is based on only the edge pixels to speed up the computation. The edge pixels are detected by intelligently scanning a small part of whole image pixels which is distributed over the image. A fast method for line and circle centre detection is also discussed. For object localization, 26 key points are defined on the soccer field. While two or more key points can be seen from the robot camera view, the three rotation angles are adjusted to achieve a precise localization of robots and other objects. If no key point is detected, the robot position is estimated according to the history of robot movement and the feedback from the motors and sensors. The experiments on NAO and RoboErectus teen-size humanoid robots show that the proposed vision system is robust and accurate under different lighting conditions and can effectively and precisely locate robots and other objects.