Mirror Field Cloud Monitoring Based On HSV Space

With the development of the society,the energy consumption is increasing year by year.The traditional fossil energy will be exhausted in the near future,and the use of fossil energy will also cause environmental pollution.Renewable energy is becoming more and more important in energy consumption because it's clean,reusable.It is very promising to use solar energy to generate electricity.As one of the most promising solar thermal power generation technologies,the solar thermal tower power system has been vigorously studied by all countries.Solar thermal tower power system is favored by people because of its simple principle and complete cleanliness.More and more solar thermal tower power plants are being planned.Monitoring clouds and forecasting the paths of the clouds are the key technologies to ensure the safe operation of solar thermal tower power system.However,the current methods in forecasting the paths of the clouds cost too much.This thesis explored the methods of monitoring clouds by combining image processing and machine vision,and achieved good effect in monitoring and tracking clouds above the heliostats.(1)Researches on the methods of image pre-processingImages taken by cameras are influenced by various elements.Noise in the images affects the quality of the images.To build the foundation for further work,it's necessary to study the methods of image pre-processing.(2)Establishment of HSV color modelHSV color model of the images of the clouds in the sky is established in this thesis.Images of the clouds in the sky are processed in the HSV color model such as enhancing the images and checking the edges of the clouds in the sky.(3)Monitoring and tracking the cloudsMoving clouds are detected using the pixel wise difference between consecutive image frames.The modeling of the background is used to separate the cloud and the sky.The improved camshift algorithm is used to achieving real-time tracking of the clouds(4)Researches on the covering of the cloudsA convolutional neural network is made in the recognition of the clouds.The speed of the moving clouds is tested based on the surf feature and image matching.The moving of the clouds is forecasted in order to judge the cover of the clouds.The study on the methods for monitoring clouds provides feasible approaches in monitoring clouds in the sky by using image processing and machine vision technology,which providing useful information for the development of the solar power tower technology.