Authors: Feilin (Falina) Lai*, Florida State University, Xiaojun Yang, Florida State University
Topics: Remote Sensing, Latin America, Urban and Regional Planning
Keywords: Informal settlements, multiple classifier systems, satellite Imagery, thematic accuracy, Latino city
Session Type: Paper
Presentation File: No File Uploaded
Informal settlements have become a global phenomenon driven by rapid urbanization and the lack of affordable formal housing. Spatial information on informal settlements is important to support various planning efforts. While satellite imagery can be an excellent source of spatiotemporal information on informal settlements, deriving such information is challenging due to the complex spectral and spatial characteristics in urban areas. In this study, we developed a multiple classifier system (MCS) to map informal settlements from remote sensor imagery. The study area encompasses the municipality of Rio de Janeiro, Brazil, which is well known for the development of informal settlements during the past several decades. We selected several robust but diverse pattern classifiers as base ones to construct different MCS structures. We found that the MCS structure can greatly affect the performance of thematic mapping.