How Does Land Use/Land Cover Map's Accuracy Depend on Number of Classification Classes?

Truong, Van Thinh; Phan, Duong Cao; Nasahara, Kenlo Nishida; Tadono, Takeo

doi:10.2151/sola.2019-006

Cited by 22 publications

(14 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…With the numerous LULC types and long-term observation, these results constituted an outstanding achievement 38 . For the detailed LULC products (Level 2), open water and mangrove had the highest levels of accuracy, accounting for over 96% in both the PA and UA.…”

Section: Resultsmentioning

confidence: 95%

First comprehensive quantification of annual land use/cover from 1990 to 2020 across mainland Vietnam

Phan

Trung

Truong

et al. 2021

Sci Rep

Self Cite

View full text Add to dashboard Cite

Extensive studies have highlighted a need for frequently consistent land cover information for interdisciplinary studies. This paper proposes a comprehensive framework for the automatic production of the first Vietnam-wide annual land use/land cover (LULC) data sets (VLUCDs) from 1990 to 2020, using available remotely sensed and inventory data. Classification accuracies ranged from 85.7 ± 1.3 to 92.0 ± 1.2% with the primary dominant LULC and 77.6 ± 1.2% to 84.7 ± 1.1% with the secondary dominant LULC. This confirmed the potential of the proposed framework for systematically long-term monitoring LULC in Vietnam. Results reveal that despite slight recoveries in 2000 and 2010, the net loss of forests (19,940 km2) mainly transformed to croplands over 30 years. Meanwhile, productive croplands were converted to urban areas, which increased approximately ten times. A threefold increase in aquaculture was a major driver of the wetland loss (1914 km2). The spatial–temporal changes varied, but the most dynamic regions were the western north, the southern centre, and the south. These findings can provide evidence-based information on formulating and implementing coherent land management policies. The explicitly spatio-temporal VLUCDs can be benchmarks for global LULC validation, and utilized for a variety of applications in the research of environmental changes towards the Sustainable Development Goals.

show abstract

Section: Resultsmentioning

confidence: 95%

First comprehensive quantification of annual land use/cover from 1990 to 2020 across mainland Vietnam

Phan

Trung

Truong

et al. 2021

Sci Rep

Self Cite

View full text Add to dashboard Cite

show abstract

“…With an increasing spatial resolution and increasing extent of Earth Observation (EO) images, the gap between historic land cover maps and current 10 m resolution products is growing (Van Thinh et al, 2019;d'Andrimont et al, 2021). This makes it difficult to identify key processes of land cover change over large areas (Veldkamp and Lambin, 2001;Vilar et al, 2019).…”

Section: Introduction 52mentioning

confidence: 99%

A spatiotemporal ensemble machine learning framework for generating land use / land cover time-series maps for Europe (2000 – 2019) based on LUCAS, CORINE and GLAD Landsat

Witjes¹,

Parente²,

Diemen³

et al. 2022

Preprint

View full text Add to dashboard Cite

A seamless spatiotemporal machine learning framework for automated prediction and analysis of long-term Land Use / Land Cover dynamics is presented. The framework includes: (1) harmonization and preprocessing of high-resolution spatial and spatiotemporal input datasets (GLAD Landsat, NPP/VIIRS) including 5 million harmonized LUCAS and CORINE Land Cover-derived training samples, (2) model building based on spatial k-fold cross-validation and hyper-parameter optimization, (3) prediction of the most probable class, class probabilities and model variance of predicted probabilities per pixel, (4) LULC change analysis on time-series of produced maps. The spatiotemporal ensemble model consists of a random forest, gradient boosted tree classifier, and an artificial neural network, with a logistic regressor as meta-learner. The results show that the most important variables for mapping LULC in Europe are: seasonal aggregates of Landsat green and near-infrared bands, multiple Landsat-derived spectral indices, long-term surface water probability, and elevation. Spatial cross-validation of the model indicates consistent performance across multiple years with overall accuracy (a weighted F1-score) of 0.49, 0.63, and 0.83 when predicting 43 (level-3), 14 (level-2), and 5 classes (level-1). The spatiotemporal model outperforms spatial models on known-year classification by 2.7% and unknown-year classification by 3.5%. Results of the accuracy assessment using 48,365 independent test samples shows 87% match with the validation points. Results of time-series analysis (time-series of LULC probabilities and NDVI images) suggest forest loss in large parts of Sweden, the Alps, and Scotland.Positive and negative trends in NDVI in general match the land degradation and land restoration classes, with “urbanization” showing the most negative NDVI trend. An advantage of using spatiotemporal ML is that the fitted model can be used to predict LULC in years that were not included in its training dataset,allowing generalization to past and future periods, e.g. to predict LULC for years prior to 2000 and beyond 2020. The generated LULC time-series data stack (ODSE-LULC), including the training points, is publicly available via the ODSE Viewer. Functions used to prepare data and run modeling are available via the eumap library for python.

show abstract

“…The large difference in the agreement reported by Vilar et al (2019) and J. Chen et al (2015) can be attributed to the number of classes in the two studies: the absolute accuracy linearly drops with the number of classes (Herold et al, 2008;Van Thinh et al, 2019), and usually the accuracy results for 6-10 classes vs 40 classes can be up to 50% better.…”

Section: Introduction 51mentioning

confidence: 84%

“…With an increasing spatial resolution and increasing extent of Earth Observation (EO) images, the gap between historic land cover maps and current 10 m resolution products is growing (d Ándrimont et al, 2021;Van Thinh et al, 2019). This makes it difficult to identify key processes of land cover change over large areas (Veldkamp & Lambin, 2001;Vilar et al, 2019).…”

Section: Introduction 51mentioning

confidence: 99%

A spatiotemporal ensemble machine learning framework for generating land use / land cover time-series maps for Europe (2000 – 2019) based on LUCAS, CORINE and GLAD Landsat

Witjes¹,

Parente²,

Diemen³

et al. 2021

Preprint

View full text Add to dashboard Cite

A seamless spatiotemporal machine learning framework for automated prediction, uncertainty assessment, and analysis of long-term LULC dynamics is presented. The framework includes: (1) harmonization and preprocessing of high-resolution spatial and spatiotemporal input datasets (GLAD Landsat, NPP/VIIRS) including 5~million harmonized LUCAS and CORINE Land Cover-derived training samples, (2) model building based on spatial k-fold cross-validation and hyper-parameter optimization, (3) prediction of the most probable class, class probabilities and uncertainty per pixel, (4) LULC change analysis on time-series of produced maps. The spatiotemporal ensemble model consists of a random forest, gradient boosted tree classifier, and a artificial neural network, with a logistic regressor as meta-learner. The results show that the most important variables for mapping LULC in Europe are: seasonal aggregates of Landsat green and near-infrared bands, multiple Landsat-derived spectral indices, long-term surface water probability, and elevation. Spatial cross-validation of the model indicates consistent performance across multiple years with overall accuracy (weighted F1-score) of 0.49, 0.63, and 0.83 when predicting 44 (level-3), 14 (level-2), and 5 classes (level-1). The spatiotemporal model outperforms spatial models on known-year classification by 2.7% and unknown-year classification by 3.5%. Results of the accuracy assessment using 48,365 independent test samples shows 87% match with the validation points. Results of time-series analysis (time-series of LULC probabilities and NDVI images) suggest forest loss in large parts of Sweden, the Alps, and Scotland. An advantage of using spatiotemporal ML is that the fitted model can be used to predict LULC in years that were not included in its training dataset, allowing generalization to past and future periods, e.g. to predict land cover for years prior to 2000 and beyond 2020. The generated land cover time-series data stack (ODSE-LULC), including the training points, is publicly available via the Open Data Science (ODS)-Europe Viewer. Functions used to prepare data and run modeling are available via the eumap library for python.

show abstract

How Does Land Use/Land Cover Map's Accuracy Depend on Number of Classification Classes?

Cited by 22 publications

References 23 publications

First comprehensive quantification of annual land use/cover from 1990 to 2020 across mainland Vietnam

First comprehensive quantification of annual land use/cover from 1990 to 2020 across mainland Vietnam

A spatiotemporal ensemble machine learning framework for generating land use / land cover time-series maps for Europe (2000 – 2019) based on LUCAS, CORINE and GLAD Landsat

A spatiotemporal ensemble machine learning framework for generating land use / land cover time-series maps for Europe (2000 – 2019) based on LUCAS, CORINE and GLAD Landsat

Contact Info

Product

Resources

About