The Study On Cloud Detection And Atmospheric Correction With New Type Satellite Data

Satellite image of earth observation is usually affected by the cloud pixel. In order to eliminate these effects, we need to do the cloud detection and reject the cloud. When the satellite receives the information from radiation on the surface of the earth, it receives atmospheric radiation information either. Only by eliminating atmospheric radiation effects can we obtain the accurate surface radiation information. Therefore, we need to do the atmospheric correction. Both FY-3 and Himawari-8 are new generation of meteorological satellite. The study on cloud detection and atmospheric correction with FY-3 MERSI and Himawari-8 AHI is not enough. This article selects the multi-channel spectrum threshold method to do the cloud detection with FY-3 MERSI data, and selects the 6s radiative transfer model to do the atmospheric correction with FY-3 MERSI data and Himawari-8 AHI data. The main work of this paper includes:(1) In this paper, we will find which channel can be used by analyzing MODIS cloud detection algorithm and spectral characteristics of FY-3 MERSI. Then we use Keams clustering analysis setting threshold on these channels direct or after normalization treatment and get a multi-channel threshold value method for MERSI. By useing this method on FY-3 MERSI map in different time different place and Contrasting with cloud detection products by satellite center, we can find that:the effect of our method is better than the cloud detection products of satellite center. This method can distinguish the cloud and snow better, have a better result on the coastline area and identifying high thick clouds.By analyzing the four seasons cloud detection results, it shows this method has some mistake on thin cloud volume and some high reflectivity area. But the overall effect is good.(2)We take the FY-3 MERSI spectral response function in 6s model and analyze the effect of parameters and its theoretical basis and get a lookup table apply to china. We use the method of 6s lookup table do atmospheric correction with satellite image data after cloud detection and getting accurate continuous surface reflectance image. By contrasting with the atmospheric correction products by satellite center, It shows that they have good consistency.(3)We use the atmospheric correction method on satellites to Japanese Himawari-8 and compare the results with before correction reflectance and the normalized vegetation index (NDVI) before or after the correction and MODIS surface reflectance products. The results show that the correction effect is good. The reflectance with band1 and band2 are reduced. The reflectance with band3 some are bigger and some are smaller.The others are increscent. NDVI is smoother, the range is broader, the peak is higher and it is closer to the actual situation. The results of AHI band1 have certain correlation with MODIS band3.