Study on Dynamic Monitoring of Urban Landscape Using Computer Vision Technology

Traditional monitoring methods are time-consuming and inefficient, and cannot meet the needs of real-time monitoring. Computer vision technology, especially remote sensing technology and street view image analysis, provides a new solution for urban landscape monitoring. By constructing a closed-loop system including data collection, processing and analysis, storage management and user interaction interface, this study realized the automatic identification and classification of urban landscape elements such as buildings, vegetation and roads. The experimental results show that the system is excellent in recognition accuracy, especially in the recognition of buildings and trees, with an accuracy of more than 90 %. In addition, the system can also capture the changes of urban landscape in real time, which provides a scientific basis for urban planning and decision-making. Although there are some influencing factors, such as lighting conditions, image occlusion and model generalization ability, the system designed in this study has demonstrated efficient, accurate and real-time monitoring ability, which has contributed to the sustainable development of cities and the improvement of residents' quality of life.