Combining Frame Rate Adaptation and Similarity Detection for Video Sensor Nodes in Wireless Multimedia Sensor Networks

Wireless Multimedia Sensor Networks (WMSNs) are composed of small embedded video sensors that allow continuous monitoring of a given territory. They collect and analyze frames from different video sensors deployed in the area of interest. One of the most important challenges in WMSN is the big data problem affecting the energy resources of the video sensors. Cameras and video-sensors send images and videos which costs in terms of memory storage, bandwidth and energy. In this paper, we propose a technique that adapts the frame rate at the level of each video-sensor. Our aim is to reduce the number of frames sent to the coordinator without losing any important information. Our approach is based on analyzing similarity between consecutive frames for each sensor. The proposed algorithm calculates the similarity by the aggregation of color and edge similarities between frames. The results of the proposed algorithm show a reduction in terms of energy consumption and sent data while guaranteeing the detection of critical events.