Scholarlink

Abstract <p>Defect inspection is critical for ensuring the safe and reliable operation of railways transportation systems. This paper presents a novel defect inspection system that combines the attention U-Net network, a type of neural network architecture, and a kind of active contour algorithm based on morphological operators to improve the accuracy of defect segmentation. The attention U-Net Network is used to generate an initial segmentation mask of the railway image with attention mechanisms that enable the network to focus on the most relevant features in the image. The active contour algorithm based on morphological operators is then applied to refine the segmentation mask. The system was tested on a dataset of railway images with various defects, and the results showed that the proposed system achieved higher accuracy in defect segmentation compared to traditional segmentation methods. The proposed system has the potential to improve the efficiency and reliability of railway defect inspection, leading to safer and more reliable railway transportation.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Attention Based U-Net Network Unified Morphological Active Contour for Accurate Defect Detection in Railways Images

  • 期刊论文

  • 发表日期：

    2025年3月1日

查看全文
  • Mohamed Ben Gharsallah,
  • Mohamed Ben Amara

摘要

Abstract

Defect inspection is critical for ensuring the safe and reliable operation of railways transportation systems. This paper presents a novel defect inspection system that combines the attention U-Net network, a type of neural network architecture, and a kind of active contour algorithm based on morphological operators to improve the accuracy of defect segmentation. The attention U-Net Network is used to generate an initial segmentation mask of the railway image with attention mechanisms that enable the network to focus on the most relevant features in the image. The active contour algorithm based on morphological operators is then applied to refine the segmentation mask. The system was tested on a dataset of railway images with various defects, and the results showed that the proposed system achieved higher accuracy in defect segmentation compared to traditional segmentation methods. The proposed system has the potential to improve the efficiency and reliability of railway defect inspection, leading to safer and more reliable railway transportation.