Lightweight Server-Assisted H-K Compression for Image-Based Embedded Wireless Sensor Network

Embedded wireless imaging systems based on wireless camera sensor network can be used to observe biological phenomena unobtrusively in various environmental monitoring applications. In our prior work, we have shown the feasibility and benefits of such a system through actual deployments. We have also shown that low-complexity data compression schemes can be employed to improve image transfer rate or to lower the energy costs of communication. In this paper, we extend upon our prior work and propose a scheme called as H-K compression, a simple lightweight image compression algorithm combining the ideas of Huffman coding and K-means clustering. Specifically, H-K compression applies K-means clustering for pixel color grouping and Huffman coding for the group color encoding, but combines the two schemes into one algorithm which can reduce the computational cost. Using 100 000 images collected from our pilot deployments at James Reserve, we study the applicability and impact of the proposed algorithm. Our results suggest that the cost of running the learning steps on an embedded sensor node may outweigh the benefit of compression, but offloading the learning can provide significant energy gains. Evaluations using the data set show that our proposed scheme compresses the data by ~57% and reduces power usage by ~43% when sending image updates from a bird nest periodically every 15 min.