Compressive sensing(CS) is an emerging technology that can recover a signal from under sampled measurements based on sparsity of the signal in some basis domain. Even though CS associated with wireless sensor networks (WSNs) has contributed in developing efficient compression and detection algorithms, most research has focused on detection problem with a simple model without considering properties of physical channels. This paper considers a compressive sensing based WSN, that exploits channel gain to transmit and detect signals efficiently. Assuming that measured signals at each sensor are correlated and sparse at some basis domain, we propose a novel sensor selection scheme and associated signaling channel design to improve detection performance. The simulation results show that the proposed method support reduction in the number of measurmenets by 6080% for a wide range of sparsity level at high and low SNRs.