Reliable Optical Camera Communication for Active Marker Identification in Camera-based Positioning Systems

BearbeiterIn:Lorenz Gorse
Titel:Reliable Optical Camera Communication for Active Marker Identification in Camera-based Positioning Systems
Typ:bachelor thesis
Betreuer:Mutschler, C.; Philippsen, M.
Status:abgeschlossen am 16. März 2018

The machine learning and information fusion group of the Fraunhofer Institute for Integrated Circuits in Nuremberg works on the next generation of optical and RF-based positioning systems. The institute operates the L.I.N.K. (german: Lokalisierung, Identifikation, Navigation, Kommunikation) test center that provides a unique evaluation and development environment for positioning systems at a size of more than 1400m2. One system in development uses an outside-in optical tracking system that locates active infrared markers via low-powered and low-cost anchors that are equipped with commodity cameras. In certain situations the system can already precisely determine the position of a marker that is in the field of view of at least two camera anchors.
Every anchor in the system extracts the pixel coordinates of all markers within its field of view. Using the intrinsic (lens parameters etc.) and extrinsic (position, angle) camera parameters each anchor derives a conceptual line in world space from the pixel coordinates of the marker. Intersections or near intersections of two lines from two cameras describe the location of the markers. With N cameras and M markers there are a total of O(N2M2) possible intersections. However, reusing information from previous frames allows us to only consider a fraction of all possible intersections.
With this approach we can continuously and reliably track markers provided they move at low speeds, even in difficult lighting conditions in a large tracking area. However, there are still three major shortcomings:

  • The system can only deliver anonymous locations. It is impossible to uniquely identify the detected markers.
  • The system may find false solutions as two lines from different markers may also generate a valid position as the lines may intersect at some (wrong) position.
  • Measurement noise that lets the system detect markers that do not exist (e.g., due to reflections or the sun) increases M and hence the number of possible intersections that need computation time.

All of these problems are gone if we could identify the markers in the camera image. Then only intersections of lines from markers with the same identification need to be computed and in addition there are no longer any false solutions. An idea from the field of Optical Camera Communication (OCC) lets a marker transmit an identification string to the anchors by encoding data into the active LED signal that it emits [2]. This has been demonstrated to work for a relatively short maximal distance between the light emitting device and the camera [1, 2, 5], with expensive high speed cameras [2], with a single photodiode to receive the signal [2, 3] (Which gives no spatial information about the signal source), or with an LCD screen to emit the signal [2, 4].
The goal of this thesis is to design, implement, and evaluate an Optical Camera Communication system that reliably transmits data across long distances under varying marker speeds and lighting conditions. The brightness of the LED should be modulated with the binary data and the receiver should recognize these variations in brightness to recover the original data from it. The marker should use this communication channel to send its binary ID to the anchors. Moreover, the system must be able to receive signals from multiple senders in parallel and must robustly handle (temporary) partial or total occlusion of senders and must account for movements of the marker (for example by means of Kalman filters). An evaluation should investigate the practical applicability of the communication system. A key metric is the bit error rate, that measures how often the receiver fails to decode the original data from the incoming brightness values.
The proposed work consists of the following:

  • Study the relevant work (publications, patents) on optical marker identification
  • Design and implement components for marker identification
  • Conceive and perform an in-depth evaluation of the system with realistic use-cases
  • Write thesis
  • Optional: Integrate the system into the existing Positioning System

The thesis must contain a detailed description of all developed and used algorithms as well as a profound result evaluation and discussion. The implemented code has to be documented and provided.

[1] T.-C. Bui and S. Kiravittaya: Demonstration of using camera communication based infrared LED for uplink in indoor visible light communication. In Proc. 6th IEEE Intl. Conf. Communications and Electronics (Ha Long, Vietnam). pp. 71-76. 2016.
[2] T. Nguyen, A. Islam, T. Hossan, and Y. M. Jang: Current Status and Performance Analysis of Optical Camera Communication Technologies for 5G Networks. In IEEE Access, vol. 5, pp. 4574–4594, 2017.
[3] P. Shamsudheen, E. Sureshkumar, and J. Chunkath: Performance Analysis of Visible Light Communication System for Free Space Optical Communication Link. Procedia Technology. Vol. 24, pp. 827–833, 2016.
[4] S. D. Perli, N. Ahmed, and D. Katabi: PixNet. In Proc. 16th Annual Intl. Conf. Mobile Computing and Networking (Chicago, Illinois, USA). pp. 137-148. 2010.
[5] S.-H. Chen and C.-W. Chow: Color-Shift Keying and Code-Division Multiple-Access Transmission for RGB-LED Visible Light Communications Using Mobile Phone Camera. In IEEE Photonics Journal. Vol. 6, no. 6. pp. 1–6. 2014.

watermark seal