Robocup @ HSRT

Hochschule Reutlingen, RT Lions

RoboCup Open Platform League: Object Detection & Recognition

Competitions @ Magdeburg, Germany | Nagoya, Japan

Group photo with all the participants (robots included) in Robocup 2017
Group photo with all the participants (robots included) in Robocup 2017

Implementation description:

Using only an image input from the camera mounted on our service robot, an object detection and recognition pipeline is built to locate objects on the cupboard and table. The object classification model is trained using MATLAB. During the competition, unknown objects are also included. Thus, using a simple similarity measure (Euclidean distance) with the extracted image features, the unknown objects are grouped with the known object classes.

Object detection pipeline for storing grocery task
Object detection pipeline for storing grocery task
Michelle Foo
Michelle Foo
PhD Student
Next
Previous