Towards large-scale mapping of local climate zones using multitemporal Sentinel 2 data and convolutional neural networks

Rosentreter, Johannes; Hagensieker, Ron; Waske, Björn

doi:10.1016/j.rse.2019.111472

Cited by 76 publications

(55 citation statements)

References 39 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Finally, our accuracies are within the range of accuracies reported from other transferability experiments using random forest classifiers 32 , 65 , acknowledging the somewhat higher accuracies using (residual) convolutional neural networks 65 – 67 . While the latter method is considered to constitute a feasible approach for automated large-scale LCZ mapping 66 , this feasibility to date has not yet been demonstrated.…”

Section: Technical Validationmentioning

confidence: 99%

Combining expert and crowd-sourced training data to map urban form and functions for the continental US

et al. 2020

View full text Add to dashboard Cite

Although continental urban areas are relatively small, they are major drivers of environmental change at local, regional and global scales. Moreover, they are especially vulnerable to these changes owing to the concentration of population and their exposure to a range of hydro-meteorological hazards, emphasizing the need for spatially detailed information on urbanized landscapes. These data need to be consistent in content and scale and provide a holistic description of urban layouts to address different user needs. Here, we map the continental United States into Local Climate Zone (LCZ) types at a 100 m spatial resolution using expert and crowd-sourced information. There are 10 urban LCZ types, each associated with a set of relevant variables such that the map represents a valuable database of urban properties. These data are benchmarked against continental-wide existing and novel geographic databases on urban form. We anticipate the dataset provided here will be useful for researchers and practitioners to assess how the configuration, size, and shape of cities impact the important human and environmental outcomes.

show abstract

Section: Technical Validationmentioning

confidence: 99%

Combining expert and crowd-sourced training data to map urban form and functions for the continental US

et al. 2020

View full text Add to dashboard Cite

show abstract

“…In case updates occur in the future, they will be tracked via the software version number and described in the changelog available on the Github Issue page. For example, some successfully tested the use of object-based image analysis (Collins and Dronova, 2019;Simanjuntak et al, 2019), others obtained promising results using (residual) convolutional neural networks (Qiu et al, 2019(Qiu et al, , 2020Yoo et al, 2019;Liu and Shi, 2020;Rosentreter et al, 2020;Zhu et al, 2020). Yet to date, the feasibility of such procedures for large-scale LCZ mapping has not yet been demonstrated (Demuzere et al, 2020a).…”

Section: Discussionmentioning

confidence: 99%

LCZ Generator: A Web Application to Create Local Climate Zone Maps

Demuzere

Kittner

Bechtel

2021

Front. Environ. Sci.

142

View full text Add to dashboard Cite

Since their introduction in 2012, Local Climate Zones (LCZs) emerged as a new standard for characterizing urban landscapes, providing a holistic classification approach that takes into account micro-scale land-cover and associated physical properties. In 2015, as part of the community-based World Urban Database and Access Portal Tools (WUDAPT) project, a protocol was developed that enables the mapping of cities into LCZs, using freely available data and software packages, yet performed on local computing facilities. The LCZ Generator described here further simplifies this process, providing an online platform that maps a city of interest into LCZs, solely expecting a valid training area file and some metadata as input. The web application (available at https://lcz-generator.rub.de) integrates the state-of-the-art of LCZ mapping, and simultaneously provides an automated accuracy assessment, training data derivatives, and a novel approach to identify suspicious training areas. As this contribution explains all front- and back-end procedures, databases, and underlying datasets in detail, it serves as the primary “User Guide” for this web application. We anticipate this development will significantly ease the workflow of researchers and practitioners interested in using the LCZ framework for a variety of urban-induced human and environmental impacts. In addition, this development will ease the accessibility and dissemination of maps and their metadata.

show abstract

“…Apart from the dataset to train the network, in the prediction phase, multisource multitemporal data fusion is also a straightforward and effective approach to further improve the obtained benchmark LCZ classification results. The effectiveness of data fusion has been shown in [18], [65], and [21]. Specifically, a final robust result can be achieved via a decision-level fusion of multiple predictions that are obtained from multisource data, such as SAR and hyperspectral image, with same or different classifiers [66]- [70].…”

Section: Discussionmentioning

confidence: 99%

Multilevel Feature Fusion-Based CNN for Local Climate Zone Classification From Sentinel-2 Images: Benchmark Results on the So2Sat LCZ42 Dataset

Qiu

Tong

Schmitt

et al. 2020

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

View full text Add to dashboard Cite

As a unique classification scheme for urban forms and functions, the local climate zone (LCZ) system provides essential general information for any studies related to urban environments, especially on a large scale. Remote sensing data-based classification approaches are the key to large-scale mapping and monitoring of LCZs. The potential of deep learning-based approaches is not yet fully explored, even though advanced convolutional neural networks (CNNs) continue to push the frontiers for various computer vision tasks. One reason is that published studies are based on different datasets, usually at a regional scale, which makes it impossible to fairly and consistently compare the potential of different CNNs for real-world scenarios. This article is based on the big So2Sat LCZ42 benchmark dataset dedicated to LCZ classification. Using this dataset, we studied a range of CNNs of varying sizes. In addition, we proposed a CNN to classify LCZs from Sentinel-2 images, Sen2LCZ-Net. Using this base network, we propose fusing multilevel features using the extended Sen2LCZ-Net-MF. With this proposed simple network architecture, and the highly competitive benchmark dataset, we obtain results that are better than those obtained by the state-of-the-art CNNs, while requiring less computation with fewer layers and parameters. Large-scale LCZ classification examples of completely unseen areas are presented, demonstrating the potential of our proposed Sen2LCZ-Net-MF

show abstract

Towards large-scale mapping of local climate zones using multitemporal Sentinel 2 data and convolutional neural networks

Cited by 76 publications

References 39 publications

Combining expert and crowd-sourced training data to map urban form and functions for the continental US

Combining expert and crowd-sourced training data to map urban form and functions for the continental US

LCZ Generator: A Web Application to Create Local Climate Zone Maps

Multilevel Feature Fusion-Based CNN for Local Climate Zone Classification From Sentinel-2 Images: Benchmark Results on the So2Sat LCZ42 Dataset

Contact Info

Product

Resources

About