Development of Ilajeland Degradation System in Crude Oil Exploitation Areas Using Satellite Imagery and Support Vector Machine Model

Abstract

Land degradation caused by crude oil exploration poses a major environmental challenge in coastal regions of Nigeria, particularly in Ilajeland, Ondo State. This study developed a land degradation detection system using remote sensing and Support Vector Machine (SVM) techniques to identify and map degraded areas within the study area. Multispectral satellite imagery obtained from Landsat 8 and Sentinel-2 sensors was preprocessed through atmospheric correction, clipping, and cloud masking to ensure data quality and consistency. Spectral indices including Normalized Difference Vegetation Index (NDVI), Soil Adjusted Vegetation Index (SAVI), Normalized Difference Water Index (NDWI), and Bare Soil Index (BSI) were extracted to characterize vegetation health and soil exposure. Due to limited ground-truth data, K-Means clustering was employed to generate pseudo-labelled training samples for supervised classification. The SVM model, implemented using the Radial Basis Function (RBF) kernel, achieved an overall classification accuracy of 92.4% with a Kappa coefficient of 0.847, indicating strong agreement between predicted and validation classes. Spatial analysis revealed severe degradation along the coastal fringe, such as in Ayetoro and Atijere, while moderate degradation was observed in inland agricultural communities. The findings demonstrate that integrating satellite remote sensing with machine learning provides an effective, scalable, and cost-efficient approach for environmental monitoring in oil-impacted regions. The developed system offers valuable support for environmental management, policy formulation, and sustainable land restoration initiatives in Ilajeland and similar coastal environments.