Innovative adaptive edge detection for noisy images using wavelet and Gaussian method

Scientific Reports, Journal Year: 2025, Volume and Issue: 15(1)

Published: Feb. 18, 2025

Edge detection is a crucial task in image processing and remote sensing, particularly for accurately identifying separating shapes noisy digital images. To enhance robustness detail edge detection, this study presents an innovative method, which integrates denoising module adaptive thresholding technique to effectively address challenges associated with Gaussian noise The proposed employs wavelet functions decompose, filter, reconstruct the image, thereby reducing impact of enhancing quality. For method based on modified OTSU utilized. Comprehensive experiments validate framework by comparing detected edges against ground truth across various levels (0.1%, 10%, 20%, 30%). median function chosen its stability convenience, while hard avoided due tendency introduce artifacts. Objective metrics, including Mean Squared Error (MSE), Accuracy, Peak Signal-to-Noise Ratio (PSNR), are employed evaluation. Comparative results indicate that outperforms traditional methods, such as Canny Roberts, showcasing effectiveness detection.

Language: Английский

Quantification of the Flood Discharge Following the 2023 Kakhovka Dam Breach Using Satellite Remote Sensing

Water Resources Research, Journal Year: 2025, Volume and Issue: 61(3)

Published: March 1, 2025

Abstract Fourteen months post the Ukrainian‐Russian war outbreak, Kakhovka Dam collapsed, leading to weeks of catastrophic flooding. Yet, scant details exist regarding reservoir draining process. By using a new technique for processing gravimetric satellite orbital observations, this study succeeded in recovering continuous changes mass with temporal resolution 2–5 days. integrating these variations imagery and altimetry data into hydrodynamic model, we derived effective width length breach subsequent 30‐day evolution discharge. Our model reveals that initial volumetric flow rate is m 3 /s, approximately 28 times average Dnipro River. After 30 days, water level had dropped by its volume was almost completely depleted km . In addition, event provides rare opportunity examine discharge coefficient—a key modeling parameter—of giant reservoirs, which find be 0.8–1.0, significantly larger than ∼0.6 value previously measured laboratory, indicating parameter may related scale. This demonstrates paradigm utilizing multiple remote sensing techniques address observational challenges posed extreme hydrological events.

Language: Английский