Author(s): Seongwook Choi; Jaehyun Shin; Dong Sop Rhee
Linked Author(s):
Keywords: No Keywords
Abstract: This research focuses on developing an image processing framework capable of identifying structural damage in hydraulic systems through advanced deep learning. By implementing the UNET model, a convolutional neural network specialized in pixel-level classification, this study successfully isolated cracked regions within complex image data. The proposed semantic segmentation approach demonstrates high levels of precision and accuracy during feature extraction, suggesting significant potential for monitoring the structural integrity of water-related facilities.
Year: 2026