자료유형 | 학위논문 |
---|---|
서명/저자사항 | Spatio-Temporal Information Extraction Under Uncertainty Using Multi-Source Data Integration and Machine Learning: Applications to Human Settlement Modelling. |
개인저자 | Uhl, Johannes Hermann. |
단체저자명 | University of Colorado at Boulder. Geography. |
발행사항 | [S.l.]: University of Colorado at Boulder., 2019. |
발행사항 | Ann Arbor: ProQuest Dissertations & Theses, 2019. |
형태사항 | 264 p. |
기본자료 저록 | Dissertations Abstracts International 81-04B. Dissertation Abstract International |
ISBN | 9781088311332 |
학위논문주기 | Thesis (Ph.D.)--University of Colorado at Boulder, 2019. |
일반주기 |
Source: Dissertations Abstracts International, Volume: 81-04, Section: B.
Advisor: Leyk, Stefan. |
이용제한사항 | This item must not be sold to any third party vendors. |
요약 | Due to advances in information and communication technology, new ways of acquisition, storage, and analysis of digital data have emerged. This constitutes new opportunities, but also imposes challenges for many scientific disciplines, including the geospatial sciences, where the availability, accessibility, and spatio-temporal granularity and coverage of environmental, geographic, and socioeconomic data is steadily increasing. Multi-source data measuring identical or related processes typically increase the reliability of knowledge derived but also lead to higher levels of discrepancies. In order to fully benefit from the value of such multi-source data, the contained information needs to be extracted effectively and efficiently, employing adequate data integration, mining, and analysis techniques. This work demonstrates how the integration of coherent multi-source geospatial data supports information extraction and analysis to generate new knowledge of both, the data itself and the underlying phenomenon, exemplified by the spatio-temporal distribution of human settlements. I present three applications in the field of human settlement modelling where data integration is a key component for knowledge acquisition. These three applications consist of i) a deep-learning based classification framework for fully automated extraction of built-up areas from historical maps in the spatial domain, ii) a machine-learning based time series classification framework for estimating changes in built-up areas in the temporal domain, based on multispectral remote sensing time series data, and iii) a novel framework for an in-depth accuracy assessment of model-generated data, exemplified by the Global Human Settlement Layer, for a detailed analysis of data uncertainty in the spatio-temporal domain, as well as across different scales and aggregation levels, attempting to quantify the fitness-for-use of such data. |
일반주제명 | Computer science. Remote sensing. Geographic information science. Geodetics. |
언어 | 영어 |
바로가기 |
: 이 자료의 원문은 한국교육학술정보원에서 제공합니다. |