For wireless communication signals, most of the traditional detection algorithms need to rely on artificial prior knowledge of the signal, which leads to the lack of adaptability of the traditional methods. A large number of experiments show that the convolutional neural network model in the research of signal recognition, compared to the traditional algorithm, not only has better detection performance, but also has a wider range of practical applications. Based on a comprehensive review of the traditional blind detection and recognition methods based on expert experience features, this paper proposes intelligent blind detection and recognition methods such as communication signal intelligent detection, intelligent modulation recognition and intelligent emitter individual recognition based on the feature autonomous learning framework of deep learning, aiming at the main problem that its feature extraction depends on human experience. The effectiveness of the algorithm is verified by simulation and measured data, and the engineering system is designed and implemented to verify the detection and recognition performance of the algorithm and system for actual communication signals.