Context-aware information processing in visual sensor network

Abstract

A visual sensor network is one of the streams of sensor network in which an image from the sensor node is transmitted to the server to be processed further without human intervention. Object Recognition in specific applications such as agriculture, defense etc, requires only the object of importance to be captured and transmitted to the processing center. This can be done by training the visual sensor node using the multilayer feed-forward technique. In the proposed work, a fruit object recognition system has been developed using the multilayer feed-forward technique of the neural network by extracting features from the sample fruit images. The experimental results reveal average recognition rate of 94.23%. © 2013 Springer-Verlag.