An efficient packet image transmission based on texture content for border side security using sensor networks

Abstract

In surveillance, several algorithms are used to extract meaningful information from an image captured. In the presence of intrusion, cameras will transmit captured images to the sink node via other intermediate nodes. Since, WSNs operate with limited resources; utilisation of resource is needed while processing and transporting images. As the node doesn't need whole image, prioritisation is one of the methods to utilise the available resource. It will prioritise images from its macro-blocks dynamically. Here the camera is attached in sensor nodes, which forms wireless multimedia sensor networks (WMSN). It applies an encoding scheme at the source node by naming the blocks as 'important' or 'not-important' based on the information they contain. Here image texture and spectral feature is used as priority measures to weight importance of macro-blocks by their GLCM properties. Experimental result discloses that priority scheme adapts itself to the quality requirements while reducing the bandwidth comparatively.