Architecture of wireless Visual Sensor Node with Region of Interest coding

Abstract

The challenges involved in designing a wireless Vision Sensor Node include the reduction in processing and communication energy consumption, in order to maximize its lifetime. This work presents an architecture for a wireless Vision Sensor Node, which consumes low processing and communication energy. The processing energy consumption is reduced by processing lightweight vision tasks on the VSN and by partitioning the vision tasks between the wireless Vision Sensor Node and the server. The communication energy consumption is reduced with Region Of Interest coding together with a suitable bi-level compression scheme. A number of different processing strategies are investigated to realize a wireless Vision Sensor Node with a low energy consumption. The investigation shows that the wireless Vision Sensor Node, using Region Of Interest coding and CCITT group4 compression technique, consumes 43 percent lower processing and communication energy as compared to the wireless Vision Sensor Node implemented without Region Of Interest coding. The proposed wireless Vision Sensor Node can achieve a lifetime of 5.4 years, with a sample period of 5 minutes by using 4 AA batteries.