Multiscale Based Characterization and Classification of Urban Land-Use

Arndt, Jacob; Lunga, Dalton; Weaver, Jeanette; LeDoux, St. Thomas Majeau; Tennille, Sarah A.

doi:10.1109/igarss.2019.8900083

Cited by 4 publications

(2 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The choice of features and their parameterization are important issues as noted in some works where certain features are shown to be better suited than others for different case studies [6], [17], [18], [38]- [42]. Another common practice is to compute local features at multiple scales or window sizes to better capture the spatial characteristics of USUs [18], [38], [40]. SVMs, classification and regression trees, and random forests have all been used to define the classifier component in these early works.…”

Section: Related Workmentioning

confidence: 99%

“…In this section, we train and test the two proposed CNN architectures and DenseNet-121 which was the overall top performing architecture from the previous experiments in Venezuela. We compare the results obtained from these architectures to a nondeep learning baseline used previously in [40]. We compare these methods for USU classification in Johannesburg, South Africa; Dar es Salaam, Tanzania; Dakar, Senegal; Addis Ababa, Ethiopia; and Nairobi, Kenya.…”

Section: B Comparisons To Engineered Features and Support Vector Machinesmentioning

confidence: 99%

See 1 more Smart Citation

Large-Scale Classification of Urban Structural Units From Remote Sensing Imagery

Arndt

Lunga

2021

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

Self Cite

View full text Add to dashboard Cite

Remote sensing in combination with deep learning has become instrumental for efficiently and accurately classifying land-use and land-cover across large geographic areas. These technologies have also been successful in characterizing urban environments in terms of their structural units, structure types, or morphological regions. In these approaches, an urban area is partitioned into regions that exhibit homogeneous physical characteristics. However, existing approaches are typically limited to a single city, use inconsistent typologies, and lack scalability and generalization capacity. In this article, we propose an urban structural units categorization scheme and demonstrate its utility by applying it to 13 cities. Inspired by the lack of scalability and generalization capacity in urban structural units mapping, we extend the reach of deep learning and conduct a set of classification experiments in all 13 cities. These experiments offer insights into the strengths and limitations of deep neural networks for classifying urban structural units over diverse geographic regions and on heterogeneous collections of satellite imagery. The efficacy of the proposed deep learning approach is compared to a baseline method of multiscale image features and support vector machines. Our validation on five cities shows that better performance is achieved with deep neural networks. Additionally, we evaluate the impact of input size, model depth, and spatial pyramid pooling to assess the generalization capacity of deep neural networks.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: B Comparisons To Engineered Features and Support Vector Machinesmentioning

confidence: 99%

Large-Scale Classification of Urban Structural Units From Remote Sensing Imagery

Arndt

Lunga

2021

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

Self Cite

View full text Add to dashboard Cite

show abstract

Urban Energy Systems: Research at Oak Ridge National Laboratory

et al. 2021

View full text Add to dashboard Cite

In the coming decades, our planet will witness unprecedented urban population growth in both established and emerging communities. The development and maintenance of urban infrastructures are highly energy-intensive. Urban areas are dictated by complex intersections among physical, engineered, and human dimensions that have significant implications for traffic congestion, emissions, and energy usage. In this chapter, we highlight recent research and development efforts at Oak Ridge National Laboratory (ORNL), the largest multipurpose science laboratory within the U.S. Department of Energy’s (DOE) national laboratory system, that characterizes the interactions between the human dynamics and critical infrastructures in conjunction with the integration of four distinct components: data, critical infrastructure models, and scalable computation and visualization, all within the context of physical and social systems. Discussions focus on four key topical themes: population and land use, sustainable mobility, the energy-water nexus, and urban resiliency, that are mutually aligned with DOE’s mission and ORNL’s signature science and technology capabilities. Using scalable computing, data visualization, and unique datasets from a variety of sources, the institute fosters innovative interdisciplinary research that integrates ORNL expertise in critical infrastructures including energy, water, transportation, and cyber, and their interactions with the human population.

show abstract