Assessment Of Rooftop Solar Energy Potential In Rural Areas Based On Deep Learning Methods

At present,solar photovoltaic building integration technology has been more mature,mainly used in urban environment at home and abroad,but combined with China’s national conditions,the building area in rural areas is far larger than that in urban areas,and there is less shelter between buildings,which is very suitable for the development of solar photovoltaic utilization system.The key to the large-scale planning of solar photovoltaic utilization system in rural areas is to effectively evaluate the photovoltaic utilization potential of building roof.The assessment of the solar energy utilization potential of the roof is mainly divided into three aspects: first,the roof area of the buildings in the research area needs to be assessed to obtain the effective roof area where the solar photovoltaic system can be installed;then,the solar energy resource potential of the target area needs to be estimated;finally,the solar energy utilization potential is assessed based on the roof area and solar radiation.The main contents of this paper are as follows:(1)Taking Wu Yang County of Henan Province as the research object,this paper first introduces the geographical and climatic environment of the research area,analyzes the feasibility of Installing Photovoltaic system,then compares and analyzes the theory of traditional image segmentation that can be used to calculate the roof area and image semantic segmentation technology based on deep learning,and finally briefly introduces the related concepts and theories of solar photovoltaic utilization.(2)Using the image semantic segmentation technology based on the deep learning method to realize the roof segmentation of low-resolution remote sensing image in rural areas.The remote sensing image of the research area is obtained by using the new earth map software,and part of the image is marked as the roof segmentation data set.The checkerboard smoothing module,multi low level feature fusion and channel attention module are integrated in the Deep Lab V3+ network to improve the segmentation accuracy of the roof area.(3)Considering the influence of clouds and atmosphere on solar radiation,based on the characteristics of the study area,the terrain and shadow analysis is carried out,and the solar radiation of the study area is estimated by SHORTWAVE-C solar radiation model.Then the solar radiation analysis tool in GIS software is used to simulate and calculate the solar radiation in the study area,and the results of shortwave-c model are compared to verify the effectiveness of the solar radiation estimation method based on SHORTWAVE-C model.(4)Based on the estimation results of available roof area and solar radiation in the study area,the potential of solar photovoltaic utilization is evaluated.Based on the installation cost of solar photovoltaic equipment,various financial subsidies and power generation income,the economic benefits of roof solar photovoltaic system are analyzed.