基于半约束条件下不透水面的遥感提取方法

Estimation of impervious surface based on semi-constrained spectral mixture analysis

不透水面作为城市生态环境的一个重要指标，被广泛应用于城市扩张监测、热岛效应分析及人类活动影响等方面的研究中。线性光谱混合模型构造简单、物理含义明确，是估算不透水面的主要方法。但是全约束的线性光谱混合模型容易在不透水面覆盖较低的地区（0-20%）出现高估，而在不透水面覆盖较高的地区（80%-100%）出现低估。因此，以黑龙江省富锦市为实验区，利用Landsat5 TM图像，讨论了线性光谱混合模型在不同端元数目和约束条件下对不透水面的估算精度，发现三端元（高反射率地物、植被及土壤）半约束条件的线性光谱混合模型估算结果最优，其均方根误差为16.71%，并结合地表温度和植被覆盖度辅助分析，去除了水田对不透水面估算的影响，提高了不透水面的估算精度。

Impervious surface plays an important role in monitoring urban sprawl and understanding human activities. Linear spectral mixture analysis （ LSMA ） is commonly used to estimate impervious surface due to its simple structure and clear physical meaning. However, previous researches found that LSMA seemed to overestimate slightly impervious surface fraction in less developed areas （ 0 -20%） but underestimate it in the central business district （ CBD） （ over 80%） . To tackle this problem, the authors developed impervious surface of Fujin Town in Heilongjiang Province from the Landsat Thematic Mapper （ TM） image by using LSMA model under different constrained conditions and end-members. The results indicated that three end-members （ high albedo, soil, and vegetation） semi-constrained LSMA provided a fine performance with a RMSE of 16. 71%. Moreover, the paddy field in impervious surface fraction image was removed by using land surface temperature and vegetation coverage data.