QR Codes for RoboCup 2012

In 2012, the RoboCupRescue Robot League will be introducing QR codes into the arena! There are several reasons for this.

  • Autonomous detection and identification of symbols is rapidly becoming practical and is something that the emergency responder community is wanting to see fielded. For instance, a future response robot might autonomously scan and map its environment for well-known signs and symbols including license plates, hazardous materials placards, building and room markings, shipping labels and other such signs, in addition to laser mapping, and record things that the operator may miss.
  • It provides a way to measure the amount of the arena that has been covered with narrow angle cameras, at a particular resolution. This year we only place a single resolution of QR code around the arena, in future years we may place a variety of resolutions. This will allow us to determine the percentage of the arena that has been inspected to different resolutions with the camera system. This complements the existing coverage figure obtained for the laser map using the round barrels.
  • It also provides a way of determining, for a larger number of points, the spatial accuracy of the map and the accuracy with which teams are able to place observed objects in the 3D environment. This encourages the more widespread use of truely integrated 3D perception using RGBD or RGBDT (T = thermal) mapping, something that only very few teams have currently implemented and yet will be very useful to the responder community.

QR codes were introduced for the first time at the 2012 German Open in Magdeburg with great success. The photograph below shows an example of a scene where environmental QR codes are present. In total 70 QR codes were scatterd around the arena.

Note that the codes may be high or low, they may be on the floor and, in the case of the raised flooring sections, they may be on the underside of platforms so robots must go underneath and look up to see them.

Points will be awarded for each QR code found. The scaling will be subject to change but as a guide, an accurately placed, correctly decoded QR code will be worth approximately 10% of an accurately placed, well identified victim.

It is very easy to incorporate QR code identification into existing image pipelines of C++ or Java using the open source ZXing library.

The C++ library can be a little bit non-obvious to obtain and use. Here is a quick, unsupported guide to get you started.

Back to the Safety, Security and Rescue Robotics Summer School page.

robocup2012/qrcodes.txt · Last modified: 2017/07/13 07:42 (external edit)
Recent changes RSS feed Donate Powered by PHP Valid XHTML 1.0 Valid CSS Driven by DokuWiki