Sensor selection for improving accuracy of target localisation in wireless visual sensor networks

Considering target localisation without prior knowledge of the target is one of the critical areas of development, in this study, the authors propose a new sensor selection solution that improves the accuracy of target localisation without prior knowledge of the target in wireless visual sensor networks. The proposed solution exploits the properties of the overlap region of the target in images: (i) the more overlap grade, the more cameras project images of the target in the same direction; (ii) the greater overlap area, the higher possibility that the target is located in the region. The authors formulate the sensor selection problem as one of maximising the utility of multiplying the overlap grade by the overlap area gained from a set of sensors. Simulation results show the effectiveness of this approach of sensor selection in terms of improving the accuracy of target localisation.