ResNet50-based satellite remote sensing application for temporary land use monitoring

Classification technology for remote sensing imagery is one of the key technologies in the application of remote sensing technology, but its development has been constrained by various factors. Due to the high spatial resolution, rich information content, complex spatial structure, and inconsistent feature distribution of remote sensing imagery, its classification work faces significant challenges, particularly in terms of classification accuracy. Traditional classification methods rely on features manually designed by experts. Determining classification features not only requires specialized knowledge but also makes it difficult to ensure the stability and accuracy of classification results, with the process being time-consuming and requiring high levels of professional expertise. In contrast, the ResNet network model demonstrates excellent performance in image classification and has gradually become the mainstream method for improving the accuracy of remote sensing image classification in deep learning network classification. Therefore, we employed the ResNet50 network to conduct classification research on remote sensing images of temporary land use in Gansu Province, and compared it with multiple classic network models: AlexNet, VGG16, GoogleNet, Vision Transformer16, DenseNet121, MobileNet, EfficientNet, SE-ResNet50, and different-level ResNet network models. This demonstrated the superiority of the ResNet50 network and the accuracy of its structure in remote sensing image classification tasks.