Mapping urban land type with multi-source geospatial big data: a case study of Shenzhen, China

Land types visually reflect the comprehensive attributes of land resources, allowing the monitoring of land development dynamics and evaluation of the land use situation. However, existing land type classification systems in urban areas lacked the integrated consideration of human characteristics and natural landscapes. Thus, this study integrated multi-source geospatial data including Weibo check-in, taxi-tracking, building-surveys, and remote sensing images to acquire anthropogenic features including building characteristics and human activities, and natural landscape features including water, vegetation, and urban vacant land. Three-level urban land type classification system was then constructed with two primary categories, six secondary categories, and 17 tertiary categories to reveal integrated characteristics of urban land resources. Based on the extracted urban areas of Shenzhen, urban land types in 2014 were mapped at street block to reveal the distribution and utilization efficiency of land resources in a visual and detailed format, thus guiding land resources optimization.

Automated summary

This study developed an integrated urban land type classification system that considered both human characteristics and natural landscapes, using multiple sources of geospatial data. The system was applied to map and analyze the distribution and efficiency of land resources in Shenzhen, providing guidance for land resource optimization.