Comparing the Accuracy of LiDAR and UAV Photogrammetry for Multi-Temporal Slope Change Detection

Accurate detection of slope deformations is essential for proactive landslide risk mitigation and geotechnical hazard management. This study compares LiDAR and UAV-based photogrammetry to evaluate their accuracy and bias in detecting surface deformation on clayey slopes. Multi-temporal point clouds and orthophotos were processed to generate co-registered digital elevation models and detect slope changes. A Gradient Boosting model was developed to quantify the relationship between LiDAR-derived and photogrammetry-derived change metrics, assess systematic biases, and predict cross-sensor discrepancies. Validation metrics used to evaluate cross-sensor performance included mean bias, root mean square error (RMSE), mean absolute error (MAE), and Lin’s concordance correlation coefficient (CCC), providing a comprehensive measure of both accuracy and precision. Results indicate that the photogrammetry consistently underestimates elevation change by approximately 0.37 m compared with LiDAR. The bias remains largely uniform across slope angles but increases with displacement magnitude. Although this study was conducted on a single clayey slope, the results indicate sources of bias that are likely to occur in other terrain types. Vegetation cover, lighting conditions, and slope steepness may further reduce photogrammetry accuracy, while LiDAR remains less affected and provides a more stable reference for long-term monitoring. The study highlights the complementary use of both methods, recommending LiDAR for periodic high-accuracy benchmarks and photogrammetry for frequent, cost-effective updates between LiDAR surveys.