Image Compression Scheam For Wireless Sensor Networks Via Tensor PCA

Wireless Multimedia Sensor Networks(WMSNs) are expected to support multimedia services such as delivery of image and video streams and have vast application prospects in environmental surveillance, mobile health care, traffic monitoring and many other fields. In addition to the general common characteristics of wireless sensor networks, the wireless multimedia sensor networks have characteristics of processing task complicated, and energy of image acquisition, processing and transmission presents "uniform" distribution. Therefore, how to reduce and balance the energy consumption of complex tasks is the key to apply widely for wireless multimedia sensor networks.In combination of the characteristic of the network architecture of Wireless Multimedia Sensor Networks, a distributed multi-node cooperative network model called DMCN is designed by using the concept of in-network processing to improve their energy, memory and computational power. To balance the energy consumption of the network, according to roles division, camera nodes and common nodes are cooperated to accomplish the workload of image acquisition, compression and transmission. Camera nodes gather images and send blocking images to the common nodes in cluster. Common nodes adaptively compress the partitioned images by using a noise-tolerant distributed image compression algorithm based on principal component analysis(PCA) called NDIC-PCA and send the compressed data to the cluster head node. Then, the cluster head node sends the compressed image data to the station. Since NDIC-PCA is used to compress grayscale images one by one, a noise-tolerant image compression algorithm based on tensor principal component analysis(TPCA) called NIC-TPCA is proposed, which can batch compress grayscale image sequence. Simulation results demonstrate that, DCNM can effectively balance the energy consumption of network and largely extend the network lifecycle. In addition, compared with previous algorithms, the proposed NDIC-PCA algorithm achieves higher peak signal to noise ratio without decreasing compression ratio. The proposed NIC-TPCA algorithm achieves high peak signal to noise ratio without decreasing compression ratio when compress image sequence in quantities and is robust to noise.