A New Cloud Motion Winds Algorithm Based On A New Generation Of Geostationary Meteorological Satellite

Wind farm data plays a decisive role in weather forecasting and natural disaster warning.However,cloud motion wind are the only wind field data in areas where stations such as oceans and deserts are scarce.At present,cloud motion wind has become an important satellite product,but its calculation method is still dominated by the largest correlation coefficient method.In recent years,satellite technology has been continuously improved,the quality of captured cloud images has been improved,time intervals have been shortened,time complexity of the maximum correlation coefficient method has increased,and the efficiency of the algorithm has been reduced.This paper describes in detail the characteristics and satellite images of the Japanese Himawari-8 satellite and the satellite imagery photographed.To use the TV-L1 optical flow model to retrieve cloud motion wind for this satellite cloud map,firstly calculate the cloud motion wind based on different solar zenith angles and cloud heights using three consecutive cloud data of different channels,and extract the details of the cloud map using the ROF model.The introduction of the TV-L1 optical flow model and the use of the pyramid algorithm to solve the problem;secondly,the height of the cloud motion wind is specified by the linear interpolation based on the standard atmospheric temperature vertical profile;then in order to improve the accuracy of the cloud motion wind and remove unreasonable wind vectors,13 types quality identification codes(QI)have been set up to carry out quality monitoring of cloud motion wind farms,and output includes latitude,longitude,wind speed,direction,cloud height,and quality identification codes.Finally,the whole cloud image data shot with Himawari-8 satellite was used to test,and the results were verified by sounding data.The results show that the TV-L1 optical flow method is suitable for cloud motion wind calculation,especially for low cloud that pressures higher than 400hpa.The TV-L1 optical flow model is not only a cloud motion wind calculation method,but also provides a new idea for prediction of satellite cloud images because the calculation result is based on a single pixel.