Distributed detection in wireless sensor networks is considered, where the decisions made by local sensors are transmitted towards a fusion center through parallel binary symmetric channels. Our contribution is to show that, when the number of local sensors approaches infinity and the channel capacity from the local sensor to the fusion center is larger than the entropy of any local decision, there exists one method in theory that the local decision can be sent to the fusion center without distortion. Correspondingly, the asymptotically optimal link bit error probability between each sensor and the FC is determined. The resulting distributed detection system can reach asymptotically the same detection performance as that with ideal communication. Although the proposed method is asymptotically optimal in nature, it is valid to the distributed detection system when the quantity of the sensors is above dozens. The simulation confirms its validity. Moreover, when compared with the ideal one, simulations confirmed that the proposed scheme achieves a significant saving of communication resource at the expense of a small reduction in the detection performance for the system with dozens of sensors.