Urban inland inundation ponding distribution rapid prediction method based on improved convolutional neural network

The invention relates to an improved convolutional neural network-based urban inland inundation ponding distribution rapid prediction method, which comprises the steps of 1, constructing a rapid prediction model data set: the data set is an inland inundation influence factor-inland inundation characterization factor data set, and the inland inundation influence factors comprise a rainfall factor, a terrain factor, a pipe network length factor and a river system distance factor; 2, building a rapid prediction model based on a convolutional neural network, carrying out training and testing, in the model building process, carrying out independent coding on different forms of waterlogging influence factors, then carrying out fusion, and finally carrying out decoding after fusion; in the training and testing process, an improved design loss function is adopted; and 3, performing model prediction and model prediction result analysis. According to the method, direct conversion of rainfall and flood elements (inundat