基于HJ-1B卫星数据的积雪面积制图算法研究

积雪是影响气候变化的重要因子, 采用更高时空分辨率的环境减灾卫星遥感数据进行积雪制图算法的研究, 对推进我国自主遥感卫星在积雪监测领域的应用具有重要意义. 采用环境减灾HJ-1B卫星数据, 以青海省果洛藏族自治州达日县为研究区, 应用归一化差值积雪指数(NDSI)法建立了基于HJ-1B卫星数据的积雪面积制图算法, 并比较MODIS与HJ-1B积雪图精度. 结果表明: 研究区HJ-1B积雪制图合理的NDSI阈值为0.37, 总分类精度达到97.97%; 与"真值"影像比较, HJ-1B积雪图K_hat系数为0.911, 高于MODIS的0.817. 说明该研究建立的基于HJ-1B积雪制图算法精度可靠, 适合对研究区积雪进行实时动态监测. HJ-1B更高的空间分辨率对提高研究区积雪覆盖面积监测精度具有重要的使用价值, 但是地形因素是影响HJ-1B数据积雪分类精度的一个重要原因, 随着坡度的增加, 分类误差也随之增大, 尤其是多测误差增加比较显著.

Snow Cover Mapping Algorithm Based on HJ-1B Satellite Data

Snow cover is an important factor affecting climate. Using the HJ satellite with higher spatial and temporal resolution for snow cover mapping has a great significance to promoting China's own remote sensing satellites in the field of snow monitoring. In this paper, the normalized difference snow index (NDSI) method based on HJ-1B satellite data is used to study the snow cover mapping algorithm in Darlag County, Qinghai Province. The accuracies of MODIS daily snow cover map and the HJ-1B snow cover map are compared. It is found that: 1) the suitable NDSI threshold of HJ-1B snow cover mapping algorithm is 0.37 and the total classification accuracy is 97.97%. 2) Compared with the MODIS daily snow cover map, the HJ-1B snow cover map has higher coherence with "true value" snow cover image with K_hat coefficient of 0.911, more than that of MODIS daily snow cover map, 0.817. The accuracy of snow cover mapping algorithm based on HJ-1B is established in this study, which is reliable for snow cover dynamic monitoring in the study area. The HJ-1B with higher spatial and temporal resolution can improve the accuracy of snow cover area monitoring. However, the terrain is an important factor for snow cover monitoring accuracy when using HJ-1B data. The snow classification error increases with slope, especially the commission error.