Open Data Cube Products Using High-Dimensional Statistics of Time Series

Roberts, Dale; Dunn, Bex; Mueller, Norman

doi:10.1109/igarss.2018.8518312

Cited by 13 publications

(22 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Twenty-five Landsat tiles in GDA94 projection from NSW, ACT and northern Victoria (each approximately 100 km × 100 km) were selected for the experiment, with 24 used for model training and validation (RGB image in Figure 1) and one (tile +14, −40 annotated with an 'X' in Figure 1) used for model testing. The DEA data cube also provided the Triple Median Absolute Deviation (TMAD), describing the median of absolute deviations from the geomedian, which provides information on annual temporal variance through three calculations of distance: Euclidean distance (edev), spectral distance (sdev) and Bray-Curtis dissimilarity (bcdev) [38].…”

Section: Input Data Preparationmentioning

confidence: 99%

Comparing CNNs and Random Forests for Landsat Image Segmentation Trained on a Large Proxy Land Cover Dataset

et al. 2022

View full text Add to dashboard Cite

Land cover mapping from satellite images has progressed from visual and statistical approaches to Random Forests (RFs) and, more recently, advanced image recognition techniques such as convolutional neural networks (CNNs). CNNs have a conceptual benefit over RFs in recognising spatial feature context, but potentially at the cost of reduced spatial detail. We tested the use of CNNs for improved land cover mapping based on Landsat data, compared with RFs, for a study area of approximately 500 km × 500 km in southeastern Australia. Landsat 8 geomedian composite surface reflectances were available for 2018. Label data were a simple nine-member land cover classification derived from reference land use mapping (Catchment Scale Land Use of Australia—CLUM), and further enhanced by using custom forest extent mapping (Forests of Australia). Experiments were undertaken testing U-Net CNN for segmentation of Landsat 8 geomedian imagery to determine the optimal combination of input Landsat 8 bands. The results were compared with those from a simple autoencoder as well as an RF model. Segmentation test results for the best performing U-Net CNN models produced an overall accuracy of 79% and weighted-mean F1 score of 77% (9 band input) or 76% (6 band input) for a simple nine-member land cover classification, compared with 73% and 68% (6 band input), respectively, for the best RF model. We conclude that U-Net CNN models can generate annual land cover maps with good accuracy from proxy training data, and can also be used for quality control or improvement of existing land cover products.

show abstract

Section: Input Data Preparationmentioning

confidence: 99%

Comparing CNNs and Random Forests for Landsat Image Segmentation Trained on a Large Proxy Land Cover Dataset

et al. 2022

View full text Add to dashboard Cite

show abstract

“…The ODC framework enables a pixel-based approach, rather than a traditional scene-based approach to analysing Landsat data, providing direct comparison of observations from specific locations acquired at two or more epochs (Dhu et al, 2017). This analytical power provides unprecedented capability for continental-scale analysis at a high temporal frequency and has been used to develop several innovative products (see Bishop-Taylor et al, 2019;Mueller et al, 2016;Roberts, Dunn, & Mueller, 2018;Roberts, Mueller, & McIntyre, 2017;Roberts et al, 2019;Sagar et al, 2017).…”

Section: Living Earth For Australiamentioning

confidence: 99%

Living Earth: Implementing national standardised land cover classification systems for Earth Observation in support of sustainable development

et al. 2021

Self Cite

View full text Add to dashboard Cite

Earth Observation (EO) has been recognised as a key data source for supporting the United Nations Sustainable Development Goals (SDGs). Advances in data availability and analytical capabilities have provided a wide range of users access to global coverage analysis-ready data (ARD). However, ARD does not provide the information required by national agencies tasked with coordinating the implementation of SDGs. Reliable, standardised, scalable mapping of land cover and its change over time and space facilitates informed decision making, providing cohesive methods for target setting and reporting of SDGs. The aim of this study was to implement a global framework for classifying land cover. The Food and Agriculture Organisation's Land Cover Classification System (FAO LCCS) provides a global land cover taxonomy suitable to comprehensively support SDG target setting and reporting. We present a fully implemented FAO LCCS optimised for EO data; Living Earth, an open-source software package that can be readily applied using existing national EO infrastructure and satellite data. We resolve several semantic challenges of LCCS for consistent EO implementation, including modifications to environmental descriptors, inter-dependency within the modular-hierarchical framework, and increased flexibility associated with limited data availability. To ensure easy adoption of Living Earth for SDG reporting, we identified key environmental descriptors to provide resource allocation recommendations for generating routinely retrieved input parameters. Living Earth provides an optimal platform for global adoption of EO4SDGs ensuring a transparent methodology that allows monitoring to be standardised for all countries.

show abstract

“…Vegetation cover, height, hydroperiod, urban and cultivated environments are the base requirements (i.e., for LCCS Level 3). However, additional datasets should be explored, such as the use of high-dimensional pixel composites/statistics of time series to distinguish urban and cultivated environments [25,30] The inclusion of datasets that can be retrieved periodically, ideally derived from the Landsat archive, would enable the robust assessment of the state of the landscape. Moreover, the retrieval of additional variables such as AGB (e.g., as derived from combinations of ALOS-1/2 PALSAR-1/2, Landsat sensor and ICESAT-1/2 or GEDI data) and water turbidity and depth (from optical sensors) would further provide information to describe evidence-based change.…”

Section: Application Within the Deamentioning

confidence: 99%

Land Cover Mapping using Digital Earth Australia

Lucas

Mueller

Siggins

et al. 2019

Data

Self Cite

View full text Add to dashboard Cite

This study establishes the use of the Earth Observation Data for Ecosystem Monitoring (EODESM) to generate land cover and change classifications based on the United Nations Food and Agriculture Organisation (FAO) Land Cover Classification System (LCCS) and environmental variables (EVs) available within, or accessible from, Geoscience Australia’s (GA) Digital Earth Australia (DEA). Classifications representing the LCCS Level 3 taxonomy (8 categories representing semi-(natural) and/or cultivated/managed vegetation or natural or artificial bare or water bodies) were generated for two time periods and across four test sites located in the Australian states of Queensland and New South Wales. This was achieved by progressively and hierarchically combining existing time-static layers relating to (a) the extent of artificial surfaces (urban, water) and agriculture and (b) annual summaries of EVs relating to the extent of vegetation (fractional cover) and water (hydroperiod, intertidal area, mangroves) generated through DEA. More detailed classifications that integrated information on, for example, forest structure (based on vegetation cover (%) and height (m); time-static for 2009) and hydroperiod (months), were subsequently produced for each time-step. The overall accuracies of the land cover classifications were dependent upon those reported for the individual input layers, with these ranging from 80% (for cultivated, urban and artificial water) to over 95% (for hydroperiod and fractional cover). The changes identified include mangrove dieback in the southeastern Gulf of Carpentaria and reduced dam water levels and an associated expansion of vegetation in Lake Ross, Burdekin. The extent of detected changes corresponded with those observed using time-series of RapidEye data (2014 to 2016; for the Gulf of Carpentaria) and Google Earth imagery (2009–2016 for Lake Ross). This use case demonstrates the capacity and a conceptual framework to implement EODESM within DEA and provides countries using the Open Data Cube (ODC) environment with the opportunity to routinely generate land cover maps from Landsat or Sentinel-1/2 data, at least annually, using a consistent and internationally recognised taxonomy.

show abstract

Open Data Cube Products Using High-Dimensional Statistics of Time Series

Cited by 13 publications

References 11 publications

Comparing CNNs and Random Forests for Landsat Image Segmentation Trained on a Large Proxy Land Cover Dataset

Comparing CNNs and Random Forests for Landsat Image Segmentation Trained on a Large Proxy Land Cover Dataset

Living Earth: Implementing national standardised land cover classification systems for Earth Observation in support of sustainable development

Land Cover Mapping using Digital Earth Australia

Contact Info

Product

Resources

About