Boundary nodes in a Wireless Sensor Network (WSN) are highly essential due to the critical functions in which they involve. Therefore, several algorithms have been proposed in the literature to detect boundary nodes. However, due to the limited power budget in the battery of sensor nodes, there is a need to develop distributed algorithms that detect boundary nodes in such ad-hoc networks with low power consumption. In this work, we propose an enhancement for a recently published algorithm for boundary nodes detection, namely Distributed Least Polar-Angle Connected Node (DLPCN), through minimizing the communication overhead. Experimental results have shown around 28\% power reduction, with preserving the great accuracy obtained by DLPCN algorithm.