Urban sound classification method based on improved deep residual network

The invention discloses an urban sound classification method based on an improved deep residual neural network. The urban sound classification method comprises the following four steps: S1, data preprocessing: carrying out graphical processing on the amplitude and frequency of sound; s2, data feature extraction: extracting audio features of the sound signals and performing data standardization processing; s3, improving a deep residual network classifier: constructing an improved deep residual network model, and sending the processed training data set into the improved deep residual network classifier for training to obtain an improved deep residual network sound classifier model; and S4, sound classification result determination: sending a sound test data set into the improved deep residual neural network sound classifier model for testing to obtain classification accuracy. Compared with a traditional classification method, the urban sound classification method based on the improved deep residual neural netwo