Meteorological data generation method based on multi-source data fusion

The invention discloses a meteorological data generation method based on multi-source data fusion, and the method comprises the steps: firstly obtaining remote sensing data and topographic data of a to-be-measured region, carrying out the preprocessing, inputting the preprocessed data into a meteorological data generation model with trained parameters, and enabling the model to finally generate the meteorological data of the region; wherein the meteorological data refers to rainfall data or evaporation data, and the remote sensing data refers to satellite remote sensing data obtained by executing corresponding meteorological tasks; the structure of the meteorological data generation model is formed by fusing a convolutional neural network and a long-short term memory network, and model parameters are obtained through training of a training data set. The method can adapt to meteorological data prediction of different topographic condition areas, so that small hydropower station multi-target optimization schedu