Planted crop structure remote sensing image classification method and device

The invention relates to a planting crop structure remote sensing image classification method and device. According to the method and the device, a primary deep learning model is adopted for prediction; extracting the boundary pixels and the internal pixels according to the prediction result; training the boundary pixels and the internal pixels respectively; predicting weights of the spectral characteristics of each boundary pixel and each internal pixel under each classification are correspondingly obtained; adding the prediction weight of the boundary pixels with the same spectral characteristics under any classification and the prediction weight of the internal pixels under the same classification; and taking the prediction weight and the highest category as prediction results of the spectral features, and constructing a deep learning prediction classification model again according to the prediction results to predict and classify the planting crop structure remote sensing images tobe classified. According