Space dust signal identification method and system based on residual error full connection network

The invention provides a space dust signal identification method and system based on a residual error full connection network, and the method comprises the steps: collecting signal data, and marking the collected signal data as a plurality of dust signals containing dust and a plurality of non-dust signals not containing dust; dividing the marked data set into a training set and a test set; constructing a residual error full-connection network model, and training the residual error full-connection network model by adopting the training set; inputting the test set into the trained residual full-connection network model, and detecting and classifying the input signals through the residual full-connection network model; if the detected signal is a dust signal, 1 is output, and if the detected signal is a non-dust signal, 0 is output. The method has relatively high identification accuracy; according to the method, the residual block and attention mechanism technology is utilized, the attention degree of the model