The Technical Study Of Clouds Belonging And Matching Based On FY-2C

In many fields,such as synoptic climatology, weather forecasting, weather modification, national economic and military, cloud plays a very important role.In the daily operational of weather forecasting,it’s a necessary measure to use satellite cloud image because of the characteristics of wide range, high time-spatial resolution. But in the meteorological operational work, forecasters only make a subjective judgment on satellite cloud image,lack of the necessary objective and quantitative for the cloud movement and changes predicting.Using Meteorological satellite remote sensing image data, matching, automatic distinguishing and tracking cloud will have a very significant significance to improving the accuracy of weather forecasting.Many domestic and foreign scholars have made different progress in the technical study of clouds matching and paths predicting on remote sensing image.On the basis of these researches,there are four steps in the research in this paper.Firstly, we make radiometric calibration and projection transformation and extract and number vector cloud using FY-2C raw data set and cloud classification products.Secondly,the clouds are belonged to two levels weather system based on the influence of large clouds, the accuracy rate of primary weather system is more than 90% and secondary weather system is more than 80%,in the results,the area of the correct division clouds account for more than 90% of the area of the whole clouds.Thirdly,select 19 image features,include the cloud area, center point of longitude and latitude and the mean, standard deviation, median and mode of clouds from four infrared remote sensing image,then use multidimensional euclidean distance method to match sequential time weather systems and clouds based on image features.The results show that the accuracy rate of weather systems is 100%, the accuracy rate of one-to-one matching clouds is more than 90%.In the study,an innovation way is that matching and path predicting is in cloud as a unit, reducing the amount of computation and more in line with daily habits;Compared with previous studies, GIS and the method of matching based on images features are introduced to the clouds matching with vector cloud. The results achieve the expected purpose and lay a good foundation for the future use in meteorological operational work.