Roof Area Recognition And PV Capacity Estimation Based On Remote Sensing Image

In order to meet the development and utilization of large-scale roof solar and PV resources in cities, many energy-management departments of developed countries have established the website of solar PV Maps, showing the sunshine of roof and shadow state, PV resources and its mode, benefit assessments, and other important data which provides public guiding services for the roof owner or roof developer. However, there is few disclosed PV Maps in China and their function is relatively simple, so it can not reach the extent for large-scale development of roof PV.If our county wants to develop large-scale roof PV, PV Maps play an important role of energy developing guidance and information services. The roof area extraction of buildings from remote sensing image is the key technology of PV Maps.With the development of remote sensing technology, the resources of high-resolution remote sensing are increasingly abundant. Because of its large amount of information, easy to get, real-time and accurate, it has been widely applied in many areas. By analyzing we find that the free satellite images of Google Earth also can satisfy the requirements. So in this paper we use the free image to extract the buildings'roof and estimate the roof area, then to estimate its PV capacity. The emphasis and difficulty of this study is building recognition and extraction from remote sensing image.In this paper, taking the dormitory buildings of Hangzhou Dianzi University as a study case, the way of building extraction from high-resolution remote sensing image is studied. This paper proposes a building extraction method based on edge analysis, by image pre-processing, threshold segmentation to remove those which do not belong to building sections, and through morphological processing, edge detection and Hough transform technique to extract buildings. Since the above method using only the edge feature, it is difficult to further develop. Therefore, this paper puts forward a method based on the combination of region and edge analysis to extract the outline of building using K-means algorithm and Otsu threshold segmentation algorithm to complete the buildings segmentation, which could remove most of non-building, and then to combine this algorithm together with edge analysis method. The result of experiment shows that the figure of the final extracted building get better, and edge position is more accurate, it indicates that this method is better than simply using edge analysis method to extract buildings' roof area.The PV installation capacity has been estimated according to the PV panel layout scheme on the available roof area which is extracted from above methods based on the remote sensing image. The estimated installation capacity value has been compared with a reference-building, and the result proves the correctness.