Abstract:Accurate measurement of water levels is essential for effectively managing reservoirs to proactively mitigate flooding and drought. Nonetheless, the inaccuracies in measurements derived from gauging station and remote sensing images impose constraints to water resource management. In this study, we developed a novel water level estimation model which utilizes solely the altitude of reliable water boundary pixels to improve the accuracy. The enhanced water boundary detection, incorporating preprocessing steps such as image filtering, resampling, and polarization multiplication, was applied to achieve sub-pixel precision in detecting water boundaries. The water boundary pixels located in layover and shadow regions, which could be misidentified due to distortion error, are eliminated based on backward geolocation. Ambiguous water boundaries, potentially indicating land with low intensity, were defined by computing their absolute derivatives, and removed. Finally, to enhance water level precision, the model computed water levels by averaging the altitudes of boundary pixels with weighting factors of local incidence angle, derivatives of detected water boundaries, and altitude distribution. Compared with the previous studies utilizing water boundaries, the proposed model demonstrated outstanding performance in improving the accuracy, up to 1/40th smaller than the spatial resolution of SAR images in Mean Absolute Error (MAE). The validation was executed over the results from >700 Sentinel-1 against the in-situ measurements obtained from multiple reservoirs and streams with significant water level fluctuations in the Korean peninsula. In this process, we found that the water boundaries located on the layover and shadow regions significantly influence the dispersion of altitude of reliable water boundary pixels. This study demonstrates the proposed model relying on remote sensing data without in-situ measurements, which holds potential applicability under situations where in-situ data are unavailable.

Robust Image-Based Water-Level Estimation Using Single-Camera Monitoring

Visual Measurement of Water Level under Complex Illumination Conditions

Method on water level ruler reading recognition based on image processing

Automatic water level monitoring system based on computer vision technology for supporting the irrigation modernization

Automatic Calibration Between Multi-Lines LiDAR and Visible Light Camera Based on Edge Refinement and Virtual Mask Matching

Field performance of the GaugeCam image-based water level measurement system

Estimating Water Levels through Smartphone-Imaged Gauges: A Comparative Analysis of ANN, DL, and CNN Models

Camera Calibration for Water-Biota Research: The Projected Area of Vegetation

Development and Deployment of a Virtual Water Gauge System Utilizing the ResNet-50 Convolutional Neural Network for Real-Time River Water Level Monitoring: A Case Study of the Keelung River in Taiwan

A Water Level Measurement Approach Based on YOLOv5s

A Complex Environmental Water-Level Detection Method Based on Improved YOLOv5m

Research on Water-Level Recognition Method Based on Image Processing and Convolutional Neural Networks

A new model for high-accuracy monitoring of water level changes via enhanced water boundary detection and reliability-based weighting averaging

A non-invasive, low-cost and easy-to-operate underwater terrain monitoring method

Water Level Retrieval Using a Posteriori Residual of GNSS Pseudorange and Carrier-Phase Observations

An intelligent water level monitoring method based on SSD algorithm

Development of Image-Based Water Level Sensor with High-Resolution and Low-Cost Using Image Processing Algorithm: Application to Outgassing Measurements from Gas-Enriched Polymer

Depth Map Reconstruction for Underwater Kinect Camera Using Inpainting and Local Image Mode Filtering

Vision-based system for the control and measurement of wastewater flow rate in sewer systems

Estimation of River Water Surface Elevation Using UAV Photogrammetry and Machine Learning

Image-Based Automatic Watermeter Reading under Challenging Environments