Liang Tiangang, Gao Xinhua, Liu Xingyuan
Key Laboratory of Grassland Agro-ecosystem, Ministry of Agriculture, College of Pastoral Agriculture Science and Technology, Lanzhou University, China.
Ying Yong Sheng Tai Xue Bao. 2004 Dec;15(12):2272-6.
In this paper, a monitoring model of snow depth was built based on the 4 scenes of NOAA satellite digital images under sunshiny condition and the corresponding ground observation data from 20 meteorological stations during 2 snow disasters from 1996 to 1997 in North Xinjiang. The pixel-based snow coverage rate and snow spatial classification were studied by using linear mixture spectrum disassembling method, and two grid data layers based quantified indices used for estimating snow hazard grade of grassland and animal husbandry were put forward. The results indicated that by using the snow monitoring model and linear mixture spectrum disassembling method, the image cell based snow depth and snow coverage rate could be calculated, and the precision of snow classification could be improved. The image cell based snow hazard index could systematically express the spatial distribution of snow, grass, animal and climate conditions, and reflect the snow hazard grade of grassland and animal husbandry.
本文基于1996 - 1997年北疆两次雪灾期间晴天条件下的4景NOAA卫星数字图像以及20个气象站的相应地面观测数据，构建了雪深监测模型。利用线性混合光谱分解方法研究了基于像元的积雪覆盖率和积雪空间分类，并提出了两个基于网格数据层的量化指标用于评估草地畜牧业雪灾等级。结果表明，利用雪灾监测模型和线性混合光谱分解方法，可以计算基于图像像元的雪深和积雪覆盖率，提高积雪分类精度。基于图像像元的雪灾指数能够系统地表达雪、草地、牲畜和气候条件的空间分布，反映草地畜牧业的雪灾等级。
