ARCAS keeps on advancing in the structure assembly by using autonomous aerial robots equipped with a robotic arm and onboard cameras.
New findings achieved within scene and objects recognition tasks cope with diverse and challenging factors that make difficult the detection from aerial robots such as lightning changes, occlusions, shadows or blurred images.
Researchers from the Institut de Robòtica i Informàtica industrial (Barcelona, Spain) have developed an effective algorithm which allows UAV detect a suitable localization for the installation of the assembly structure, but also recognize the bar-shaped structures in real time and with high accuracy during the insertion tasks. This is a useful feature that let onboard cameras zooming into specific parts of the scene or object facilitating the grasping labour.
This is the first set of experiments towards the perception for aerial and space manipulation of ARCAS system.
For more details, you can download the deliverable Del 3.1 Scene and object recognition from the link below.
You can also state tuned following @ARCAS-FP7 in Twitter