Utilizing UAV and orthophoto data with bathymetric LiDAR in google earth engine for coastal cliff degradation assessment
Abstract This study introduces a novel methodology for estimating and analysing coastal cliff degradation, using machine learning and remote sensing data. Degradation refers to both natural abrasive processes and damage to coastal reinforcement structures caused by natural events. We utilized orthop...
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Nature Portfolio
2025-01-01
|
| Series: | Scientific Reports |
| Subjects: | |
| Online Access: | https://doi.org/10.1038/s41598-024-84404-1 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Abstract This study introduces a novel methodology for estimating and analysing coastal cliff degradation, using machine learning and remote sensing data. Degradation refers to both natural abrasive processes and damage to coastal reinforcement structures caused by natural events. We utilized orthophotos and LiDAR data in green and near-infrared wavelengths to identify zones impacted by storms and extreme weather events that initiated mass movement processes. Our approach included change detection analysis to estimate eroded areas. Next, by applying Random Forest classifier within Google Earth Engine, we evaluated the importance of features in detecting these degraded zones. We tested the algorithm’s performance using datasets of varying resolutions (10 cm, 20 cm, 50 cm, and 100 cm), and a UAV dataset acquired two years later to validate results. The classifier achieved an overall accuracy of approximately 90% across all datasets. The findings indicate that DEM products in green and near-infrared wavelengths are similarly important, while reflectance maps and orthophotos suggest that red and near-infrared wavelengths play a significant role in identifying degradation. These results suggest that it is feasible to monitor coastal degradation caused by natural disasters using diverse sensors within a single training framework. |
|---|---|
| ISSN: | 2045-2322 |