Estimation of roof photovoltaic potential and carbon reduction potential based on semantic segmentation method

Abstract

Since industrialization, the global surface temperature has been on an upward trend, and so far it has had a serious impact on the global ecosystem and socio-economic environment. Therefore, it is necessary to reduce carbon emissions through global joint efforts, maintain the stability of the surface temperature not exceeding 2°C, and achieve global carbon neutrality as soon as possible. Solar energy is the cleanest renewable energy and has good prospects for sustainable development in the future. Installing solar photovoltaic (PV) systems on the roofs of buildings has become the most extensive method of utilizing solar energy. In this study, we used the semantic segmentation method to estimate the solar photovoltaic potential based on identifying roofs from remote sensing images, and then combined the photovoltaic module parameters derived from building features with solar radiation data to evaluate the solar photovoltaic potential. We did experiments in Haizhou Economic Development Zone, Lianyungang City, Jiangsu Province. The results show that the number of available roofs of the photovoltaic system is 671, the available roof area is 925,810 square meters, and the annual photovoltaic power generation potential of the study area is 119,300.97 MW$\cdot$h/year, which has a large solar photovoltaic potential. The method to perform precise calculation of specific rooftop solar PV potential developed in this study will guide the formulation of energy policy for solar PV in the future.