Super-large visual field surface quarantine method and device based on convolutional neural network

The invention discloses an ultra-large visual field surface quarantine method and device based on a convolutional neural network, and the method comprises the steps: carrying out the deep learning modeling of a convolutional neural network model through a single-stage target detection algorithm; expanding a small target detection layer, and preprocessing an input image; photo segmentation: before model reasoning, the input photo is segmented, and the segmentation scheme comprises 1: 4 segmentation, 1: 9 segmentation and 1: 16 segmentation; model application: identifying the segmented sub-images; and manual confirmation and correction are carried out, and personnel carry out real-time checking and confirmation through a tablet computer or a PC, and carry out manual correction on wrong labels. According to the super-large-view surface quarantine method and device based on the convolutional neural network, under the cooperation of a wharf port machine, the technical means of hanging machine high-definition photo