Abstract:

The issue of informal settlements in post-apartheid South Africa remains severe, constituting a significant barrier to urbanization and social equity. This study taking Cape Town as a case example, uses high-resolution remote sensing imagery and geographic spatial datasets to identify the spatial distribution of informal settlements in Cape Town using feature-based method, maximum-minimum distance with building features method, and random forest with building features method, and analyzes the impact of roads, railways, schools, and hospitals on the distribution of informal settlements. The results show that albedo and building density are key features for identifying informal settlements, the feature-based method and random forest with building features method showing similar results of settlement areas. The feature-based method identifies settlements in a more block-like pattern, the maximum-minimum distance with building features method produces blurred boundaries, the edge identified by random forest with building features method shows positional deviation. Transportation infrastructure plays a dominant role in the distribution of informal settlements, with few areas having access to all facilities simultaneously. This study provides a scientific method for identifying the spatial distribution of informal settlements, offering insights for urban morphology evolution and urban planning in Africa.

Key words: Cape Town, informal settlements, feature-based method, maximum-minimum distance method, random forest method, urban facilities

摘要：

后种族隔离时代，南非的非正式聚落问题依然严峻，成为制约城市化和社会公平发展的重要障碍。本研究以南非开普敦为例，使用高分辨率遥感影像和地理空间数据集，运用特征法、最大最小距离叠加建筑特征法和随机森林叠加建筑特征法，识别并绘制开普敦的非正式聚落，分析公路、铁路、学校和医院对非正式聚落分布的影响。研究结果表明，反照率和建筑密度是非正式聚落的关键特征，特征法和随机森林叠加建筑特征法识别的面积较为接近。特征法识别的聚落分布较为块状，最大最小距离叠加建筑特征法识别的结果边界模糊，随机森林叠加建筑特征法识别结果的边缘存在位置偏差。交通设施在非正式聚落分布中具有主导作用，但聚落内几乎没有可同时通达四种设施的区域。研究为非正式聚落的空间分布提供了科学的识别方法，可为非洲城市形态演化、城市规划等提供一定的借鉴。

关键词: 开普敦, 非正式聚落, 特征法, 最大最小距离法, 随机森林法, 城市设施