Distributed detection is studied and implemented mainly for the multiple sensors detection considering such factors as the bandwidth, time-delay, computation capability, etc. It is featured that the binary hypothesis test with which the fusion center makes a decision is performed based on the local observation of a single sensor. In a single station, the factors restricting the centralized detection are overcome by applying the high performance communication and the server, where the original observation of a single sensor without any loss can be transmitted to the fusion center to improve the detection performance. Based on the minimal risk rule, the theories of distributed detection and centralized detection are analyzed. The features and difficulties in implementation of both strategies are discussed. The fundamental factors in engineering of the centralized detection are studied combined with the multiple radars detection in the single station.