A Lightweight Algorithm for Dynamic Target Acquisition and Transmission
Ke-Jiang Xiao, Rui Wang, Ji Wen and Bo-Wei Chen
Due to the limited capacity and cost of a sensor node, its computation resource and energy supply will be overloaded when the node is allocated a heavy-duty task, such as processing, storing or transferring large amount of image data. For dynamic target, there are a lot of useless images sensed by image sensor nodes and these images do not have complete feature information of the monitoring target. If these images are processed, stored and transferred, there is nothing to be gained for the system identification performance and some unnecessary energy consumption will be caused. For the above problem, we propose a light-weight image processing and transmission method on dynamic target called LIAT. First, to decrease the large amount images’ requirement for the storage in the process of computing, LIAT reuse one image storage space to store the useless images. Second, to resolve the problem of continuous images real-time processing caused by dynamics of the monitoring target, we propose an adaptive real-time detection method for image integration detection to increase the velocity of image integration detection via part image processing and simple threshold method. Finally, to reduce the communication energy consumption, LIAT only sends the integration images to the base station via complete judgment of the dynamic images sensed by image sensor nodes. We conduct both indoor and outdoor environmental experiments. The results demonstrate that although LIAT needs a small amount of extra computing resource cost, it can significantly reduce the storage and transmission consumption while guaranteeing the perceived performance.
Keywords: Sensor networks; dynamic target; image acquisition; storage consumption; transmission energy consumption.