Pyramid neural network image defogging method and system based on end-to-end multi-size fusion

The invention discloses a pyramid neural network image defogging method and system based on end-to-end multi-size fusion. Five groups of feature maps of a foggy image in different sizes and sub-regions are extracted by using a backbone network in an image defogging model; performing feature enhancement on the five groups of feature maps by using a feature pyramid network structure in the image defogging model to obtain five groups of feature maps after feature enhancement; fusing the five groups of feature maps through a spatial multi-size feature superposition fusion method to obtain fused features; a decoder in the image defogging model is used for further fusing and decoding the fused features to obtain intermediate estimation parameters of the network; and reconstructing the network intermediate estimation parameter and an original foggy image input by the network by using a physical recovery module to obtain a fogless image. According to the method, the local features and the global features can be fused