Files in this item



application/pdfXU-DISSERTATION-2020.pdf (53MB)
(no description provided)PDF


Title:Cybergis-enabled remote sensing data analytics for deep learning of landscape patterns and dynamics
Author(s):Xu, Zewei
Director of Research:Wang, Shaowen
Doctoral Committee Chair(s):Wang, Shaowen
Doctoral Committee Member(s):Guan, Kaiyu; Jiang, Zhe; Rhoads, Bruce; Usery, E. Lynn
Department / Program:Geography & Geographic InfoSci
Degree Granting Institution:University of Illinois at Urbana-Champaign
Deep learning
Landscape dynamics
Abstract:Mapping landscape patterns and dynamics is essential to various scientific domains and many practical applications. The availability of large-scale and high-resolution light detection and ranging (LiDAR) remote sensing data provides tremendous opportunities to unveil complex landscape patterns and better understand landscape dynamics from a 3D perspective. LiDAR data have been applied to diverse remote sensing applications where large-scale landscape mapping is among the most important topics. While researchers have used LiDAR for understanding landscape patterns and dynamics in many fields, to fully reap the benefits and potential of LiDAR is increasingly dependent on advanced cyberGIS and deep learning approaches. In this context, the central goal of this dissertation is to develop a suite of innovative cyberGIS-enabled deep-learning frameworks for combining LiDAR and optical remote sensing data to analyze landscape patterns and dynamics with four interrelated studies. The first study demonstrates a high-accuracy land-cover mapping method by integrating 3D information from LiDAR with multi-temporal remote sensing data using a 3D deep-learning model. The second study combines a point-based classification algorithm and an object-oriented change detection strategy for urban building change detection using deep learning. The third study develops a deep learning model for accurate hydrological streamline detection using LiDAR, which has paved a new way of harnessing LiDAR data to map landscape patterns and dynamics at unprecedented computational and spatiotemporal scales. The fourth study resolves computational challenges in handling remote sensing big data and deep learning of landscape feature extraction and classification through a cutting-edge cyberGIS approach.
Issue Date:2020-07-15
Rights Information:Copyright 2020 Zewei Xu
Date Available in IDEALS:2020-10-07
Date Deposited:2020-08

This item appears in the following Collection(s)

Item Statistics