Construction of Urban Greenway Network System Based on Remote Sensing Image and VGG

Urban greenways, as an important component connecting urban ecological networks, are of great significance in improving the quality of urban ecological environment and enhancing the well-being of residents. This article first reviews the development process of greenways, analyzes the planning strategies and implementation paths of existing greenway network systems, and explores the benefits of greenways in optimizing urban spatial structure and promoting healthy lifestyles. On this basis, this article proposes an urban greenway network system that combines remote sensing images and visual geometry group (VGG). By preprocessing, feature extraction, and intelligent image recognition of remote sensing image data, efficient and automated monitoring of urban greenway networks is achieved. The research methods include the acquisition and preprocessing of remote sensing image data, the construction of intelligent image recognition models, model training and parameter optimization, system architecture design, and system integration and functional testing. This article selects the VGG model as the basis and quickly adapts to the recognition task of urban greenway network systems through transfer learning. Comparative experiments are conducted with the residual network (ResNet) and Inception. The experimental results show that the VGG model outperforms the ResNet and Inception models in key indicators such as overall classification accuracy, Kappa coefficient, and node detection accuracy. The overall classification accuracy reaches a maximum of 98.2%; the Kappa coefficient reaches a maximum of 0.98; the node detection accuracy reaches a maximum of 84.5%. The final results indicate that intelligent image recognition technology based on VGG can effectively improve the recognition accuracy and management efficiency of urban greenway network systems.