Reliable environmental monitoring of critical infrastructure such as industrial pipelines, offshore platforms, and chemical processing facilities is essential for preventing environmental hazards and ensuring operational safety. However, many existing monitoring systems rely on single-sensor modalities that suffer from significant performance limitations when operating under variable environmental conditions such as low illumination, fog, or atmospheric disturbances.To address these limitations, this study proposes a hybrid sensor intelligence framework that integrates multispectral optical sensing and 360° LiDAR measurements within a probabilistic robotics architecture. The proposed framework employs a Bayesian inference model supported by particle filtering to fuse heterogeneous sensor observations and estimate the likelihood of environmental anomalies such as chemical leakage, corrosion indicators, and structural deformation. By combining complementary sensing capabilities with uncertainty-aware decision-making, the framework improves the robustness of anomaly detection in complex monitoring environments.The methodology was validated using a controlled infrastructure monitoring testbed that simulated realistic environmental conditions, including variable lighting, fog interference, and dust disturbances, while introducing representative anomalies such as simulated chemical leakage and structural displacement.Experimental results demonstrate that the proposed hybrid monitoring system significantly outperforms standalone sensing approaches. The hybrid model achieved an overall detection accuracy of 95.3%, compared with 83.4% for LiDAR-only systems and 81.7% for optical-only systems. Additionally, the probabilistic fusion framework reduced the false positive rate to 3.1%, representing a reduction of approximately 40–60% relative to single-sensor configurations. The system also maintained high reliability under conditions where individual sensors experienced performance degradation, demonstrating strong robustness to environmental disturbances.These findings indicate that integrating optical and LiDAR sensing within a probabilistic fusion framework provides a powerful foundation for reliable anomaly detection in autonomous environmental monitoring systems. The proposed hybrid architecture offers a promising pathway toward scalable, trustworthy, and fully autonomous monitoring solutions for critical remote infrastructure.
Keywords: LiDAR-Based sensing, autonomous environmental monitoring, bayesian sensor fusion, cyber-physical monitoring systems, hybrid sensor fusion, infrastructure anomaly detection, intelligent infrastructure monitoring, multispectral optical imaging, particle filter algorithms, probabilistic robotics
