Building Monitoring by Remote Sensing and Analysis of Distributed Photovoltaic Construction Potentials

Abstract

Distributed photovoltaic is a typical type of clean energy and has the characteristics of small investment and fast construction. Distributed photovoltaic power can address the power consumption problem in rural areas with energy shortage and industrial areas with high load densities. Building roofs are important carriers for distributed photovoltaic facilities and thus directly related to the construction potential of distributed photovoltaic. Therefore, monitoring and analyzing the spatial distribution of buildings in China has an important reference value for the planning and construction of distributed photovoltaic. In this research, we use 2 m-resolution satellite remote-sensing images as the data source and extract the building zones in China using the deep learning technology. The areas of building roofs in typical regions are calculated according to the proportion coefficients of these regions. Subsequently, we analyze the spatial characteristics of the building roofs in China and the distribution pattern for construction potentials of distributed photovoltaic. Based on this, we propose some suggestions for the construction path of distributed photovoltaic in different areas considering the spatial distribution of population. The research shows that the accuracy of building extraction by remote sensing was 81.63%, which can satisfy subsequent data analysis requirement. Approximately 1.4×104 km2 of building roofs in China have the potential to develop distributed photovoltaic. We suggest that distributed photovoltaic should be developed hierarchically following the principle of local power generation and local consumption. The provinces in China can be categorized into four echelons and the construction should start from the east part of China that is densely-populated and has great potentials for distributed photovoltaic development. Moreover, it is necessary to establish a nation-wide dynamic monitoring mechanism for distributed photovoltaic construction based on satellite remote sensing, thereby supporting the dynamic update of the distributed photovoltaic planning path.

Keywords

• satellite remote sensing，building extraction，distributed photovoltaic，deep learning