An Agricultural Greenhouse Extraction Method Based on Sentinel-2 Data

We proposed a novel approach combining Convolutional Neural Networks (CNN) with transfer learning, utilizing a ResNet34 network pre-trained on ImageNet to achieve high-precision identification of agricultural greenhouses using Sentinel-2 high-resolution imagery.