Low-overhead image compression in WMSN for post disaster situation analysis

Abstract

Due to inherent limitations in terms of processing power, storage, and bandwidth, data processing is a challenge in wireless multimedia sensor network (WMSN). Transmitting raw data is very costly while limited processing power of the sensor nodes prevents sophisticated multimedia processing. Exploring low-overhead data compression technique is a solution towards this problem. In this work we propose an energy saving image compression technique for WMSN by partial discrete cosine transform (PDCT) considering the application of post-disaster situation analysis through image capturing of the affected area. Here the DCT coefficients of only the last two rows of macroblock of the sensory image are propagated thereby saves energy by transmitting reduced number of bits. Also an existing tree based routing scheme is modified to employ in-network aggregation at merging nodes along the tree saving energy further. Finally the design feasibility along with simulation results including statistical analysis is presented to evaluate efficacy of the scheme in terms of two conflicting parameters viz. energy consumption and peak signal to noise ratio (PSNR). The comparative results confirm our scheme's supremacy in WMSN application domain over existing methods.