Scholarlink

<p>In response to growing vulnerabilities to anthropogenic activities and climate change, understanding human and environment interactions in coastal megacities has become critical. This study examines land use and land cover (LULC) changes in Lagos, Nigeria, between 2017 and 2024 as a representative coastal megacity in the Global South. While previous research has assessed LULC dynamics in Lagos, this study provides new information using the AlphaEarth satellite embedding dataset by comparing the performance of four machine learning classifiers: Random Forest, Gradient Boosted Trees, Support Vector Machine, and Classification and Regression Tree. The dataset was provided by Google DeepMind and is accessible via the Google Earth Engine (GEE) catalogue. A total of 2,400 manually collected samples were used for training (70%) and validation (30%) of the models in GEE using the hold-out method. The six land cover classes used are built-up, water, vegetation, agriculture, wetlands, and bare land. Across the four classifiers, findings revealed increases in built-up areas (5.4–15.9%) and agricultural land (5.0–30.1%), coupled with declines in vegetation (11.4–19.0%) and wetlands (1.9–28.2%) between 2017 and 2024, which highlight urban expansion, deforestation, and wetland encroachment. All classifiers achieved strong performance, with overall accuracies exceeding 93%, and the RF classifier outperforming others with accuracy above 97%. These outcomes provide insights on methodological differences in LULC detection in Lagos, with significant implications for sustainable urban planning, resource management and climate resilience.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Machine Learning-Based Land Use Land Cover Change Detection Utilising the AlphaEarth Satellite Embedding Dataset in Lagos Nigeria (2017–2024)

  • 期刊论文

  • 发表日期：

    2026年7月22日

查看全文
  • Oluwafemi M. Odunsi,
  • Olabisi S. Obaitor,
  • Margaret Y. Odunsi,
  • Valerie Graw,
  • Andreas Rienow

摘要

In response to growing vulnerabilities to anthropogenic activities and climate change, understanding human and environment interactions in coastal megacities has become critical. This study examines land use and land cover (LULC) changes in Lagos, Nigeria, between 2017 and 2024 as a representative coastal megacity in the Global South. While previous research has assessed LULC dynamics in Lagos, this study provides new information using the AlphaEarth satellite embedding dataset by comparing the performance of four machine learning classifiers: Random Forest, Gradient Boosted Trees, Support Vector Machine, and Classification and Regression Tree. The dataset was provided by Google DeepMind and is accessible via the Google Earth Engine (GEE) catalogue. A total of 2,400 manually collected samples were used for training (70%) and validation (30%) of the models in GEE using the hold-out method. The six land cover classes used are built-up, water, vegetation, agriculture, wetlands, and bare land. Across the four classifiers, findings revealed increases in built-up areas (5.4–15.9%) and agricultural land (5.0–30.1%), coupled with declines in vegetation (11.4–19.0%) and wetlands (1.9–28.2%) between 2017 and 2024, which highlight urban expansion, deforestation, and wetland encroachment. All classifiers achieved strong performance, with overall accuracies exceeding 93%, and the RF classifier outperforming others with accuracy above 97%. These outcomes provide insights on methodological differences in LULC detection in Lagos, with significant implications for sustainable urban planning, resource management and climate resilience.