Street scene image semantic segmentation method based on cross aggregation model

The invention discloses a street scene image semantic segmentation method based on a cross aggregation model, and the method comprises the steps: dividing all image data in a Cityscapes data set into a training set, a verification set and a test set, and carrying out the data enhancement and preprocessing of all image data in the training set; constructing a ResNet50 network model by using a residual structure to obtain four feature maps; the method comprises the following steps: fusing two groups of non-adjacent feature maps by using two cross aggregation modules, then realizing re-fusion of the feature maps by using an integration module, finally selecting features beneficial to segmentation by using a feature selection unit, and outputting the feature maps; and finally, outputting a semantic segmentation result. According to the method, the problem that multi-scale objects and local region categories are mixed and are difficult to segment in semantic segmentation in the prior art is solved. 本发明公开了一种基于交叉聚合模