Subpixel Snow Mapping Of Tibetan Plateau AVHRR/2 Data With The Third Band Reflectance

The monitor of snow area’s variety has important research and social implications,there are many advantages in using medium and low resolution satellite remote sensing data to monitor the change of snowy in a large area,but because there are so many kinds of features on the earth’s surface and the inherent resolution issues of the data,the application of remote sensing data to snow monitoring is often influenced by mixed pixels.Based on the remote sensing characteristics of snow,and aiming at the problem of large errors in the binary classification algorithm commonly used in snow remote sensing monitoring,on the basis of considering the geographical environment characteristics of the Qinghai-Tibet Plateau,this study develops a hybrid pixel decomposition algorithm using medium resolution satellite remote sensing data and a method of extracting endmembers using pure pixel index,taking AVHRR/2,the medium-spatial resolution images,as the data source of this study,and using the existing band information of the data to establish a reliable mid-infrared radiation transfer equation.The reflectance information of the mid-infrared band used in the remote sensing data used in the successful reverse performance research area,and then the multi-terminal element linear hybrid spectral decomposition model is applied,and the snow terminal elements and non-snow terminal elements in the cells are extracted based on the pure pixel index method,and the snow area abundance in a single cell is automatically and efficiently obtained,and on the basis of the above classification results,based on the spatial dependence of the ground objects,the sub-cell-level snow cover map exceeding the resolution of the remote sensing image itself is realized by cyclic iteration method.The results show that the proposed mid-infrared radiation transmission equation has guiding significance for the reflectivity information of the infrared band in the AVHRR/2 data,and the extraction of this information can add a new mapping basis for the historical snow filling map of AVHRR/2 data in the Qinghai-Tibet Plateau,and the multi-terminal element linear hybrid pixel decomposition model used in the study can effectively achieve the purpose of dismixing,which is of great significance for the snow accumulation information filling map of the sub-cell scale of AVHRR/2 data on the Qinghai-Tibet Plateau.