Design Of Sun Tracking Controller For Solar Panel Based On Image Processing

From the perspective of alleviating the shortage of non renewable energy and building a green ecological civilization society,the development of solar energy resources has been regarded as an important measure to adjust the energy structure of many countries and reduce the pollutants generated by fuel combustion.China is a large country with solar radiation resources,but photovoltaic power generation still has low efficiency and high cost,so how to maximize the solar radiation received per unit area and improve the utilization of solar radiation energy has important significance and scientific value.In this paper,the sun position model,forecast and identifying model of weather and spot image center positioning algorithm are studied,and the design scheme of the sun tracking control device with the forecast and identifying model of weather and feedback of image processing is proposed.The main research details are as follows:1.A sun tracing method based on feedback of image processing with forecast and identifying model of weather is proposed.By simplifying the selection of sun position model and sun imaging mode,the overall system layout of sun tracing controller is determined.At the same time,the neutral gray light blocking plate is installed outside the lens,which not only solves the problem of direct viewing light of camera,but also ensures maximum protection in multiple weather conditions Leave the original information of the facula image.2.The weather prediction model of the sun-tracking controller is studied,and the simplified processed weather data is used as input.The generalized RBF network was trained by the algorithm that initializes network parameters by self-organizing determining center method and training parameters by gradient descent,and the training model for the first four days of the day was established.The accuracy of the prediction results reached 88.89%,At the same time,combined with the prediction probability and the output current of photovoltaic panel,the fuzzy recognizer is established to infer the current weather conditions and improve the performance of the sun tracking controller to cope with the weather impact.3.According to the wavelength difference between the light emitted by the sun as a light source and the light reflected by the background object in the image,the difference between the sun object and the background object in the hue component in the image is very obvious.According to his color model,the hue component can be extracted and the spot can be marked well.At the same time,the processing algorithm of sunshine hue component is studied.Through the hue component extraction,denoising,binarization,contour extraction,sun object center location and filtering and other processing methods,the coordinate information of the sun center position in the image is obtained,and the center location deviation is less than 5 pixels.4.Based on TMS320F28335,the hardware and software implementation of the terminal of the sun tracking controller is proposed.The actuator and motor of the PTZ are selected and the monitoring platform interface of the sun tracking controller is built.After analyzing the test results of the sun tracking controller,it can not drive the PTZ in rainy weather,which saves the power consumption of the device itself.In the cloudy and sunny days with spot images,the average deviation of sunshine orientation and altitude angle is less than 0.19 °.The solar panel with sun tracking function has a 38.4% increase in solar radiation conversion rate compared with the static installation,which ensures the accuracy of sun tracking and enhances the adaptability to rainy and rainy weather conditions.The paper has 49 figures,9 tables,and 79 references.