jismart2024057: Research of Mountain-sea City Characteristics Based on Semantic Segmentation Technology: A Case Study of Qingdao

Abstract

As a typical coastal city, Qingdao's urban development has been profoundly shaped by its coastal spaces, which play a pivotal role in defining the city's landscape. To better understand the characteristics of these spaces, this study integrates image semantic segmentation, hierarchical clustering, and field investigations to analyze 28 representative urban coastal areas, including May Fourth Square, the Third Bathing Beach, and Xiaoqingdao Island. Using the Mask2Former model, we quantitatively decomposed and calculated the proportions of various landscape elements within these spaces, and applied hierarchical clustering to categorize the landscape types. The results reveal six major landscape types, highlighting the visual and functional diversity of Qingdao's coastal areas. This research provides valuable scientific insights and data that can inform the planning, design, and management of coastal mountain landscapes in future projects, laying a foundation for sustainable urban development.