Assessing Land Cover Changes Using the LUCAS Database and Sentinel Imagery: A Comparative Analysis of Accuracy Metrics DOI Creative Commons
B. Hejmanowska,

Piotr Kramarczyk

Applied Sciences, Год журнала: 2024, Номер 15(1), С. 240 - 240

Опубликована: Дек. 30, 2024

Classification of remote sensing images using machine learning models requires a large amount training data. Collecting this data is both labor-intensive and time-consuming. In study, the effectiveness pre-existing reference on land cover gathered as part Land Use–Land Cover Area Frame Survey (LUCAS) database Copernicus program was analyzed. The classification carried out in Google Earth Engine (GEE) Sentinel-2 that were specially prepared to account for phenological development plants. performed SVM, RF, CART algorithms GEE, with an in-depth accuracy analysis conducted custom tool. Attention given reliability different metrics, particular focus widely used (ML) metric “accuracy”, which should not be compared commonly “overall accuracy”, due potential significant artificial inflation accuracy. LUCAS 2018 at Level-1 detail estimated 86%. Using updated dataset, best result achieved RF method, 83%. An overestimation approximately 10% observed when reporting average ACC ML instead overall OA metric.

Язык: Английский

Assessing carbon storage dynamics in an ecological civilization demonstration zone amid rapid urbanization: A multi-scenario study of Guizhou Province, China DOI Creative Commons
Rui Chen, Xuehai Fei,

Jingyu Zhu

и другие.

Resources Environment and Sustainability, Год журнала: 2025, Номер unknown, С. 100223 - 100223

Опубликована: Апрель 1, 2025

Язык: Английский

Процитировано

0
Legacy of severe soil degradation hinders the buildup of mineral-associated soil organic carbon DOI Creative Commons
Otávio dos Anjos Leal, Rüdiger Reichel,

Holger Wissel

и другие.

The Science of The Total Environment, Год журнала: 2025, Номер 978, С. 179445 - 179445

Опубликована: Апрель 19, 2025

Язык: Английский

Процитировано

0
Assessing Land Cover Changes Using the LUCAS Database and Sentinel Imagery: A Comparative Analysis of Accuracy Metrics DOI Creative Commons
B. Hejmanowska,

Piotr Kramarczyk

Applied Sciences, Год журнала: 2024, Номер 15(1), С. 240 - 240

Опубликована: Дек. 30, 2024

Classification of remote sensing images using machine learning models requires a large amount training data. Collecting this data is both labor-intensive and time-consuming. In study, the effectiveness pre-existing reference on land cover gathered as part Land Use–Land Cover Area Frame Survey (LUCAS) database Copernicus program was analyzed. The classification carried out in Google Earth Engine (GEE) Sentinel-2 that were specially prepared to account for phenological development plants. performed SVM, RF, CART algorithms GEE, with an in-depth accuracy analysis conducted custom tool. Attention given reliability different metrics, particular focus widely used (ML) metric “accuracy”, which should not be compared commonly “overall accuracy”, due potential significant artificial inflation accuracy. LUCAS 2018 at Level-1 detail estimated 86%. Using updated dataset, best result achieved RF method, 83%. An overestimation approximately 10% observed when reporting average ACC ML instead overall OA metric.

Язык: Английский

Процитировано

2