PSI - Issue 70

Available online at www.sciencedirect.com

ScienceDirect

Procedia Structural Integrity 70 (2025) 447–452

Structural Integrity and Interactions of Materials in Civil Engineering Structures (SIIMCES-2025) Corrosion Assessment Using Image Based Deep Learning Models Bagathi HarshaVardhan a , Mallika Alapati a, *, Alapati Rahul a , Lemuel Sathwik a

a Department of Civil Engineering, VNR Vignana Jyothi Institute of Engineering and Technology, Hyderabad, Telangana, India.

Abstract Steel components like rebars in concrete, structural steel, steel pipes, wind turbines etc. are prone to corrosion in aggressive environments, posing significant maintenance challenges. Visual inspection plays important role in asset management that greatly benefits from automation. Using artificial intelligence to assist inspections can enhance safety, reduce access costs, provide objective classification, and integrate with digital asset management systems. With the advancements in deep learning and image processing, corrosion detection using Convolutional Neural Networks (CNNs) presents an efficient alternative to manual inspection The study presented utilized deep learning Convolutional Neural Network model ResNet for corrosion detection. To achieve human-level accuracy, the training of a deep learning model requires a large dataset and intensive image labeling. In this work, a dataset of 300 images labeled as three classes i.e, ‘No corrosion’, ‘Medium corrosion’, ‘Severe corrosion’ is used. A compara tive evaluation with the CNN architectures, namely ResNet, and VggNet was performed. The implementation details, including data preprocessing, pretrained architectures, and evaluation are analyzed and presented. © 2025 The Authors. Published by ELSEVIER B.V. This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0) Peer-review under the responsibility of International Conference on Structural Integrity Organizers

Keywords: Corrosion detection, Image Processing, Deep Learning, Convolutional Neural Networks, ResNet, and VggNet.

1. Introduction: Corrosion is a major challenge when it comes to infrastructure maintenance, especially in steel structures exposed to harsh environmental conditions. According to Bastian et al. (2001), traditional inspection techniques like manual visual assessments and non-destructive testing (NDT) are labor-intensive, time-consuming, and prone to human error. With the advancements in artificial intelligence (AI) and computer vision, deep learning-based approaches have come out as powerful tools for automating corrosion detection and classification (Cha et al., 2018).

* Corresponding author. Tel.: +91-8179722294. E-mail address: mallika_a@vnrvjiet.in

2452-3216 © 2025 The Authors. Published by ELSEVIER B.V. This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0) Peer-review under the responsibility of International Conference on Structural Integrity Organizers 10.1016/j.prostr.2025.07.076