Study On The Method Of Detecting The Centroid Location Of Solar In Complex Weather

To review the history of human being, we can find that the development of society greatly depends on the energy. The energy consumption is increasing with the rapid productivity growth, while the traditional energy especially the non-renewable energy sources, such as coal, gas, oil etc., are facing extinction. Furthermore, the emissions, which from the traditional sources burning, keep growing that make the environment getting worse. These problems will bring significant crisis to humankind in future development, and then the new green energy gains its popularity. As a new energy, the solar power has huge deposits and is good for environment. No matter we transfer it to heat or electricity, the solar power has great potential in future development and utilization. However, there are some disadvantages about solar energy, such as low density of radiation, alternating in day and night, and irradiation angle change with time and so on. These problems have negatively impact on the efficiency of photoelectric conversion. Especially the reducing of the vertical radiation energy, due to irradiation angle change, will lead the photoelectric conversion of the PV array decline significantly. With the widely use of concentrated photovoltaic system, the requirement of the irradiation angle accuracy is more higher than before. The efficiency and stability of transferring would raise up, if we can increase the accuracy of the sun’s center position detection, especially under complex weather.In this paper we make a systematic study on the method of the sun’s center position detection under complex weather, and the main research work is as follows:1. We analyze the development of solar detection and tracking device, and elaborate the advantages and the disadvantages of the application of solar energy. We also analyze the relationship between the image sensing device and precision of the sun’s center position detection. We get the required pixel and size of the image sensing device which can obtain the positioning accuracy of the detection reach up to 0.001°.2. We design a pro-processing method based on heat kernel and local variance. By using this approach, the image noise is reduced and the structural information of the image is kept, and the dynamic range of local image is extended without over enhancement.3. We design a segmentation method based on heat kernel. We initialize the graph edge weight set from the heat kernel matrix of the image, and then use improved Min-Cut/Max-Flow method to cut the graph. At last we obtain high quality results.4. We analyze several edge detection algorithms and centroid extraction algorithms, and propose a modified morphological algorithm, combined with the modified weighted centroid method to extract the position of the center of the solar image. At last, we prove the feasibility of the proposed algorithm through the experiments of the simulation.