The Centroid Position Detection Algorithm Based On Image Processing Research

With the development of human society, energy plays an important role inour life. Coal, oil, natural gas and other non-renewable energy sources are facingextinction. The energy crisis is the most important issue for us in the future.Solar energy is a new green energy which is storage rich, clean and no pollution.We can make use of solar energy to transfer it to other form of energy which wealways used, like heat and electricity, as we should find the new non-pollutingenergy sources to replace the old one, so we can save the earth from energycrisis. However, solar energy has some disadvantages, like low density, sunlightalternating in day and night and the changing of light intensity over time. If thepower generation system and the sunlight angle appear the deviation, the outputpower of the PV array will decline due to the reducing of the vertical radiationenergy. Therefore, locating the sun trajectory efficiently, detecting the sun’scenter position accurately and feeding the changes between solar elevation angleand azimuth in real-time can improve the utilization of the device of energytransferring and maximize the use of solar energy.This paper design the overall program about the detection of the sun’s centerposition, specific works are as follows：1、This paper research the advantages and the disadvantages of theapplication of solar energy, and analyze the development of solar detection andtracking device in the future. Then decide to choose the3million-pixelhigh-precision CCD image sensing device to detect the center position of the sunin the image. The largest imaging pixel is2048*1536. At last, the positioningaccuracy of the detection of the sun is reached up to0.001°.2、This paper propose an improved Canny edge detection algorithm. Byimproving an algorithm the paper designs an adaptive filtering scheme,improving the ability of eliminating the noise to avoid the loss of importantinformation of the sun image. Meanwhile, the algorithm can automatically setthe level of the threshold. Not only short the running time of the algorithm, butalso reduce the possibility of phenomenon of false edges.3、The paper analyze a lot of algorithms of the centroid extraction and decide to improve least squares fitting circle algorithm to extract the position ofthe center of the mass solar image. And based on the data to calculate elevationand azimuth angle of the sun in real-time. The experiments of the simulationprove the feasibility of the proposed algorithm at last.4、This paper design a GUI interface and make use of the CCD camera tocollecting the sun’s image in a real-time through software programming, andconnect to the computer by the USB in order to extract the continuous imageframe, Wiener filtering, iterative threshold value method, an improved Cannyedge detection operator and an improved the circle fitting Least Squaresalgorithm are used to achieve even greater precision and faster speed to locatethe position of the center of the sun.