Knowledge-based Classification of Land Use and Landform for Disaster Risk Assessment

对地面状况的了解是控制自然灾害造成的损失的关键因素。然而，要收集大范围地面状况的信息并不是一件容易的事。为了解决这一问题，本研究主要从以下几个方面进行了探讨：土地利用分类将古代地图数字化的土地利用数据与现代卫星数据相结合，在地理信息系统（GIS）上建立了土地利用数据库。在本研究中，卫星图像分类是基于图像对象而不是单个像素。地貌分类基于土地利用及其变化，运用推理规则进行地貌分类。这些推理规则是两种分类之间的关系，是基于地貌学知识构建的。表层土壤结构收集了大量钻孔测井资料，建立了钻孔测井资料数据库。在此基础上结合地形分类，选取典型土壤剖面进行地形分类。这组土壤剖面可以看作是本研究目标区域（千叶市）表层土壤结构的代表。基于该表土模型进行了地震风险评价。此外，还进行了一系列的微震测量和地震观测。根据这些研究，发现地面的动力特性和地震风险受到当地地形和土壤条件的强烈影响。

The knowledge about the condition of the ground is a crucial factor that controls damage due to natural disasters. However, to collect information on the ground condition over a wide area is not an easy task. In order to solve this issue, the following topics have been discussed in this research :1. Land use classificationBy combining land use data digitized from maps in the old times and those obtained from satellite data in the present time, a land use database has been constructed on a geographical information system (GIS). In this study, the satellite image classification is based on image objects rather than individual pixels.2. Landform classificationBased on the land use and its change, landform classification has been obtained by applying inference rules. These inference rules are the relationship between the two classifications and was constructed based on the knowledge about geomorphology.3. Surface soil structureA number of boring logs have been collected to construct a boring log database. Based on this database together with the landform classification, typical soil profiles have been selected for each landform classification. This set of soil profiles can be considered to represent a surface soil structure in the target area in this research (Chiba city).4. Seismic risk assessmentBased on this surface soil model, seismic risk assessment has been conducted. In addition, a series of microtremor measurement and also earthquake observation have been carried out. According to these studies, it was found that dynamic properties of the ground and hence the seismic risk are strongly affected by local land form and soil conditions.