Scholarlink

Under the widespread adoption of solar photovoltaic (PV), this study utilized a deep neural network model combined with 9 key indices to identify and evaluate PV deployment in Guangzhou, China. By incorporating the advanced U2-NET semantic segmentation model, high-accuracy identification of existing PV systems from high-resolution satellite imagery was achieved. This method effectively captured PV systems’ coordinates and area information. The general patterns and influential factors of the spatial distribution of PV systems were evaluated, confirming the feasibility and efficiency of this technical process in practical applications. The results revealed that the spatial distributions of both rooftop and ground PV systems were closely correlated to three specific key indicators: distance from roads (DFR), slope (SLO), and population density (PEO). This study not only revealed the distribution of the existing PV systems in Guangzhou but could also provide information for scientific planning and implementation of future PV projects, promoting the decarbonization of urban regions.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Evaluation of PV System Distribution in Guangzhou Using U2-Net Semantic Segmentation Model and Satellite Image

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Junyi Yang,
  • Lihua Zhao,
  • Haoshan Ren

摘要

Under the widespread adoption of solar photovoltaic (PV), this study utilized a deep neural network model combined with 9 key indices to identify and evaluate PV deployment in Guangzhou, China. By incorporating the advanced U2-NET semantic segmentation model, high-accuracy identification of existing PV systems from high-resolution satellite imagery was achieved. This method effectively captured PV systems’ coordinates and area information. The general patterns and influential factors of the spatial distribution of PV systems were evaluated, confirming the feasibility and efficiency of this technical process in practical applications. The results revealed that the spatial distributions of both rooftop and ground PV systems were closely correlated to three specific key indicators: distance from roads (DFR), slope (SLO), and population density (PEO). This study not only revealed the distribution of the existing PV systems in Guangzhou but could also provide information for scientific planning and implementation of future PV projects, promoting the decarbonization of urban regions.