天山中段的山地针叶林带很大程度上影响了该地区整体卫星雪盖的识别精度,多角度卫星遥感技术的发展为林区积雪识别提供了新的途径.本文选取了2000年4月至2001年6月,10个时段研究区内无云覆盖的(Multi-angle ImagingSpectroRadiometer)MISR多角度数据,首先对红光波段不同角度观测结果组成的角度谱图像进行非监督分类,以确定天山林带的分布区域,然后在玛纳斯河中下游与那拉提山东部选取典型像元,分析这些像元红光波段各角度反射率在林区不同积雪覆盖状况下的表现差异.研究发现,若林区存在积雪,0°,±26.1°,±45.6°五个观测角度反射率的平均值大于0.1,在部分降雪月份,后向45.6°观测的反射率大于天顶方向观测的2.5倍.根据这一结论,给出基于MISR数据的研究区不同时段的积雪识别结果.结果表明,MISR红光波段对林区积雪反应敏感,不同角度观测的反射率在林区有雪和无雪时差异较大,故可利用多角度遥感信息进行林区积雪识别.

A forest zone extending from west to east of Tianshan Mountain greatly affects the snow mapping accuracy in this area. The ability of the Multi-angle Imaging SpectroRadiometer (MISR) to obtain simultaneous measurements at nine view angles has provided unique information of forest. In this paper, we provided a new method to retrieve snow information in coniferous forest of Tianshan Mountain from MISR data. First, we chose 10 MISR scenes ranging from April 2000 to July 2001, and then ISODATA unsupervised classif ication was applied to the reflectance image viewed at f ive different view angles (0°, ±26.1°, ±45.6°) in red band acquired in summer to get the forest-region information. According to the angular signature of two study sites under certain conditions of snow distribution, we found that, when there was snow in the forest, the average of f ive view angles of red band was bigger than 0.1 or the reflectance viewed at ?45.6° was 2.5 times more than the reflectance viewed at nadir. Results indicate that directional effects are associated with the snow conditions in the forest, especially in the red band, and the surface anisotropy signature has signif icant potential in retrieving snow information in forest regions.