Application of Multi-Source Remote Sensing Data Fusion in Monitoring the Spatial Pattern Changes of Oil and Gas Field Environment

Oil and gas extraction activities play a crucial role in meeting global demands but often come with significant environmental consequences. These operations have the potential to change landforms, deplete water supplies, disturb natural ecosystems, and discharge pollutants into the soil, water, and air. Land deterioration, soil erosion, vegetation loss, along with surface and groundwater pollution, are some of the environmental impacts linked to oil and gas operations. This research investigates the application of multi-source remote sensing data fusion to monitor and analyze spatial pattern changes in oil and gas field environments. Efficient Pigeon Inspired Optimized Spatiotemporal Graph Neural Networks (EPIO-SGNN) to predict the land cover changes of oil and gas field environments from 2015 to 2023. Multi-source remote sensing data were collected from Landsat-7 satellite imagery, Shuttle Radar Topography Mission (SRTM), and Digital Elevation Models (DEMs). The data is pre-processed to remove sensor or atmospheric noise that degrades image quality and normalizes the data. The pre-processed data are extracted by using principal component analysis (PCA) to fuse the data, and feature-level fusion is performed. The result demonstrates that the proposed method has achieved high accuracy rates in spatial prediction of land cover dynamics. The performance metrics are accuracy (98%), recall (98.06%), F1 score (98.31%), and precision (98.2%). Findings indicate significant changes in land use, including an increase in land affected by industrial activities and a reduction in natural vegetation cover, highlighting the environmental consequences of oil and gas extraction. Forecasts for 2030 and 2045 point to ongoing trends in habitat change, highlighting the significance of sustainable management techniques. It provides a replicable framework for environmental monitoring in oil and gas fields worldwide, highlighting the importance of multi-source data fusion approaches in capturing intricate spatial and temporal trends in complex environments.