中文摘要：

人口易损性研究的目的在于鉴别出易损的人群并为灾害管理者提供决策依据。然而，在以往的地质灾害人口易损性评估中，所用数据几乎全来自统计数据，其评估结果精度往往只能达到次一级的行政区划。对于难以获得统计数据的区域，人口易损性评估难以进行。鉴于此，本文提出基于遥感技术的解决方案。该方案包括图像融合、图像解译、人口易损性模型构建、人口易损性计算等主要内容。在人口易损性评估中，人口分布是一个必要且重要的因素。为了获得真实有效的人口数据，利用统计模型方法进行人口估算，考虑到人口活动的时空分布特征，提出“基于土地利用的人口活动强度指数”作为新指标替代原有的人口数学密度反映人口分布。其它两个参与评估的指标包括距离医院的距离和年龄结构。应用指数加权方法，本文对广州市萝岗区进行了人口易损性评估，结果显示低、较低、中、较高、高人口易损区分别占萝岗区总面积的61．80％、25．29％、5．18％、5．21％和2．51％。

英文摘要：

The goal of population vulnerability study is to identify the vulnerable groups and provide decision makers with scientific basis for their DPM （disaster prevention and mitigation） decisions. However, in previous studies on population vulnerability assessment, nearly all the input data were from statistics and the assessment resolution could only reach the second level of the administrative area. And to the areas at lower administrative level which are lacking in statistical data, the traditional method is unusable. To resolve this problem, an innovative solution based on remote sensing is brought forward in this paper, which includes image fuse, image interpretation, population vulnerability modeling, and population vulnerability calculation. In population vulnerability assessment, the most important indicator is population density. To get more accurate and detailed population density data, a statistical model is established to estimate the population on village level. Then, on considering that human activities are related to land use, a new index of ＂intensity of human activities based on land use＂ which can reflect the population distribution more realistically is created to replace the traditional arithmetic population density. The other two indicators are age struc~re and the distance to hospital. The weighted average method of index is applied to assess the population vulnerability of Luogang District of Guangzhou. The result shows that the very low, low, moderate, high and very high vulnerability areas account for 61.80%, 25.29%, 5.18%, 5.21% and 2.51% of the total area of the district, respectively.