We propose an artificial neural network-embedded visible light positioning (VLP) scheme based on a conventional monitoring system in which the transmitter and receiver comprise a single illuminating LED and two standard surveillance cameras, respectively. Experimental results show that this simple system achieved millimeter-level positioning accuracy and exhibited good antitilt ability and robustness. The VLP system achieved high-precision three-dimensional positioning with a mean positioning error (MPE) of 0.76 cm, even if the object was inclined at different angles. Moreover, the MPEs of the single-light system were only 0.65 and 0.67 cm when the tilt angle was 15 deg and −15 deg, respectively. This scheme is a low-cost solution for accurate positioning that could be applied in marine vessel operations, mining operations, shopping center, supermarket, factories, and warehouses where monitoring systems presently exist. |
ACCESS THE FULL ARTICLE
No SPIE Account? Create one
CITATIONS
Cited by 4 scholarly publications.
Light emitting diodes
Cameras
Artificial neural networks
Imaging systems
Neural networks
Neurons
Optical engineering