Urban landscape modeling and algorithms under machine learning and remote sensing data

Traditional urban landscape modeling relies on limited geographic information data and sensor observation data, but urban landscapes are dynamic and cannot generate accurate urban landscape models. By utilizing remote sensing data and machine learning technology to capture the spatiotemporal dynamics of urban landscapes, the accuracy of urban landscape modeling is improved. This article collected sufficient urban remote sensing images and temporal data, preprocessed the collected data, and used Residual Network (ResNet) feature extractors to analyze remote sensing image data. It integrated the output of the ResNet feature extractor with urban temporal data and inputs it into the Long Short-Term Memory (LSTM) model. This article constructed the ResNet LSTM model. The results from the test set indicated that the ResNet LSTM model had an average accuracy of 97.0% for urban landscape classification. The ResNet LSTM model can effectively improve the accuracy of urban landscape classification and provide an effective method for accurately generating urban landscape models.