Fog, freezing rain, and snow (melt) quickly condense on road surfaces, forming black ice that is difficult to identify and causes major accidents on highways. As a countermeasure to prevent icing car accidents, it is necessary to predict the amount and location of black ice. This study advanced previous models through machine learning and multi-sensor-verified results. Using spatial (hill shade, river system, bridge, and highway) and meteorological (air temperature, cloudiness, vapour pressure, wind speed, precipitation, snow cover, specific heat, latent heat, and solar radiation energy) data from the study area (Suncheon–Wanju Highway in Gurye-gun, Jeollanam-do, South Korea), the amount and location of black ice were modelled based on system dynamics to predict black ice and then simulated with a geographic information system in units of square metres. The intermediate factors calculated as input factors were road temperature and road moisture, modelled using a deep neural network (DNN) and numerical methods. Considering the results of the DNN, the root mean square error was improved by 148.6 % and reliability by 11.43 % compared to a previous study (linear regression). Based on the model results, multiple sensors were buried at four selected points in the study area. The model was compared with sensor data and verified with the upper-tailed test (with a significance level of 0.05) and fast Fourier transform (freezing does not occur when frequency = 0.00001 Hz). Results of the verified simulation can provide valuable data for government agencies like road traffic authorities to prevent traffic accidents caused by black ice.

To prevent collapse accidents at construction sites, the marker-based displacement measurement method was developed. However, it has difficulty in obtaining accurate measurements at long distances (>50 m) in an outdoor environment because of camera movements. To overcome this problem, marker-based structural displacement measurement models using image matching and anomaly detection were designed in this study. Then, the performance of each model in terms of camera movement error correction was verified through comparison with that of a conventional model. The results show that the systematic errors due to camera movements (<1.7°) were corrected. The detection rate of markers with displacement reached 95%, and the probability that the error size would be less than 10 mm was ≥ 95% with a 95% confidence interval at a distance of more than 100 m. Moreover, the normalized mean square error was less than 0.1. The models developed in this study can measure the pure displacement of an object without the systematic errors caused by camera movements. Furthermore, these models can be used to measure the displacements of distant structures using closed-circuit television cameras and markers in an outdoor environment with high accuracy.

Flood risk in small streams is rising under climate change, as small catchments are highly vulnerable to short, intense storms. We develop a high-resolution assessment that integrates a Digital Surface Model (DSM), a Digital Elevation Model (DEM), and airborne LiDAR within a MATLAB (2025b) hydraulic workflow. A hybrid elevation model uses the DEM as baseline and selectively retains DSM-derived structures (levees, bridges, embankments), while filtering vegetation via DSM–DEM differencing with a 1.0 m threshold and a 2-pixel kernel. We simulate 10-, 30-, 50-, 100-, and 200-year return periods and calibrate the 200-year case to the July 2025 Sancheong event (793.5 mm over 105 h; peak 100 mm h⁻¹). The hybrid approach improves predictions over DEM-only runs, capturing localized depth increases of 1.5–2.0 m behind embankments and reducing false positives in vegetated areas by 12–18% relative to raw DSM use. Multi-frequency maps show progressive expansion of inundation; in the 100-year scenario, 68% of the inundated area exceeds 2.0 m depth, while 0–1.0 m zones comprise only 13% of the footprint. Unlike previous DSM–DEM studies, this work introduces a selective integration approach that distinguishes structural and vegetative features to improve the physical realism of small-stream flood modeling. This transferable framework supports climate adaptation, emergency response planning, and sustainable watershed management in small-stream basins.

This study aims to improve the accuracy of bathymetry predicted by gravity-geologic method (GGM) using the optimal machine learning model selected from machine learning techniques. In this study, several machine learning techniques were utilized to determine the optimal model from the performance of depth and gravity anomalies. In addition, a tuning density contrast calculated from satellite altimetry-derived free-air gravity anomalies (FAGAs) was applied to estimate enhanced bathymetry. By comparison with shipborne depth, the accuracy of the bathymetry estimated by using satellite altimetry-derived FAGAs and machine learning was evaluated. The findings reveal that the bathymetry predicted by the optimal machine learning using the Gaussian process regression and the GGM with a tuning density contrast can enhance the accuracy of 82.64 m, showing an improvement of 67.40% in the RMSE at shipborne depth measurements. Although the tuning density is larger than 1.67 g/cm³, bathymetry using satellite altimetry-derived FAGAs and machine learning can be effectively improved with higher accuracy.