Several protocols and algorithms are used for energy efficient routing in Wireless Sensor Networks (WSNs). These protocols enable not only to reduce communication latency in these networks but also to maximize the network lifetime that is directly related to energy sensors due to the constraint of the batteries. Some of these protocols use clustering or avoiding holes to reach this goal. We have proposed, in this paper, a new hybrid algorithm based on clustering and convolution filters for efficient routing in terms of energy. For this reason, we present the network as a grayscale image. Our algorithm happens in two steps, each of these steps uses a different convolution filter. One kernel filter is used to determine the nodes that participate in the routing by using the mean filter. Another kernel is to pass packets from a source to the destination by using the gradient to select the routes over the most efficient battery.
