Pattern Recognition of Grating Perimeter Intrusion Behavior in Deep Learning Method

An intrusion behavior recognition method based on deep learning is proposed in this paper order to improve the accuracy of raster perimeter behavior. The Mach–Zehnder fiber optic interferometer was used collect external vibration signal sensing unit, capture signal, use cross-correlation characteristic obtain minimum frame length and preprocess according strength. signals were superimposed several sections intercepted by fixed window length; spectrum information obtained Fourier transform stationary signals. convolution neural network introduced into pattern optical defense zone, different characteristics extracted, so as realize accurate identification Experimental results showed that highly sensitive events, could effectively reduce false alarm rate signals, efficiency recognition.