Lightning arrester body identification and segmentation system and method based on convolutional neural network

The invention discloses a lightning arrester body identification and segmentation system based on a convolutional neural network. A graph preprocessing module of the lightning arrester body identification and segmentation system preprocesses a historical lightning arrester infrared image; the data expansion module is used for expanding the preprocessed infrared images of the lightning arrester; a lightning arrester labeling module performs pixel-level labeling on the lightning arrester body in the expanded data set sample; the network parameter setting module sets network learning parameters of the convolutional neural network model; the network training module inputs the labeled data set into a convolutional neural network model with network learning parameters for training to obtain a trained convolutional neural network model; the lightning arrester body recognition and segmentation module transmits a lightning arrester infrared image collected in real time to the trained convolutional neural network model