Distributed barcode tracking in visual sensor networks
Adrian Munteanu, Eli De Poorter, Leander Hendrikx · 2017
A novel approach for robot tracking and identification based on barcodes is proposed in this paper. The proposed system tracks robots fitted with barcodes identifying them. The system performs distributed visual processing and collaborative barcode tracking, whereby the nodes exchange processed visual information with appropriate neighboring nodes, informing them about incoming targets. The proposed tracking system is deployed on a visual sensor network (VSN) based on the Raspberry Pi. Its practical realization includes various components that handle real-time collaborative tracking, streaming, communication, configuration, management, monitoring, and deployment. The proposed system is able to efficiently and accurately track barcodes in real time. Complexity modeling demonstrates that the proposed distributed system brings substantial complexity reductions compared to a system of independently operating cameras. Experimental results demonstrate that the proposed system can track multiple barcodes in real time with an accuracy of less than 1 centimeter.