Abstract:In this study, a faster region-based convolutional neural network is constructed and applied to the combined vision and thermographic images for automated detection and classification of surface and subsurface corrosion in steel bridges. First, a hybrid imaging system is developed for the seamless integration of vision and infrared images. Herein, a three-dimensional red/green/blue vision image is obtained with a vision camera, and a one-dimensional active infrared (IR) amplitude image is obtained from the infrared camera for temperature measurements with halogen lamps as the heat source. Subsequently, the three-dimensional red/green/blue vision image is converted to a two-dimensional chroma blue- and red-difference (CbCr) image because the CbCr image is known to be more sensitive to surface corrosion than the red/green/blue image. A combined three-dimensional (CbCr-IR) image is then constructed by fusing the two-dimensional CbCr image and the one-dimensional infrared image. For the automated corrosion detection and classification, a faster region-based convolutional neural network is constructed and trained using the combined three-dimensional CbCr-IR images of surface and subsurface corrosion on steel bridge structures. Finally, the performance of the trained, faster region-based convolutional neural network is evaluated using the images acquired from real bridges and compared with faster region-based convolutional neural networks trained by other vision and IR-based images. The uniqueness of this study is attributed to the (1) corrosion detection reliability improvements based on the fusion of vision and infrared images, (2) automated corrosion detection and classification with a faster region-based convolutional neural network, (3) detection of subsurface corrosion that is not detectable using vision images only, and (4) application to field bridge inspection.

Imaging-Based Rating for Corrosion States of Weathering Steel Using Wavelet Transform and Pso-Svm Techniques

Corrosion Assessment of Carbon Steel Using Texture and Color Features

Evaluation of corrosion morphology on mechanical properties of reinforcing steel bars using numerical simulation and simplified analytical model

Data-driven Corrosion Failure Probabilistic Assessment Model for Reinforced Concrete Structures

Corrosion grade classification: a machine learning approach

Structural Health Monitoring Data Reconstruction of a Concrete Cable-Stayed Bridge Based on Wavelet Multi-Resolution Analysis and Support Vector Machine

Probabilistic fatigue life assessment of bridge welded joints considering steel corrosion

Damage identification for simply-supported bridge based on SVM optimized by PSO (PSO-SVM)

Particle Swarm Optimization-Based SVM for Classification of Cable Surface Defects of the Cable-Stayed Bridges

Study on performance evaluation and maintenance management system of weathering steel bridge

Recognition of Corrosion State of Water Pipe Inner Wall Based on SMA-SVM under RF Feature Selection

Study of Evaluation and Management System for Weathering Steel Bridges

Segmentation and grade evaluation of corrosion on hydraulic steel gates based on image-level labels

Image Processing-Based Pitting Corrosion Detection Using Metaheuristic Optimized Multilevel Image Thresholding and Machine-Learning Approaches

A Hybrid Model Combining Wavelet Transform and Recursive Feature Elimination for Running State Evaluation of Heat-Resistant Steel Using Laser-Induced Breakdown Spectroscopy

Steel bridge corrosion inspection with combined vision and thermographic images

Comprehensive Safety Evaluation of Corroded Circular Steel Tubes under Compression Based on Image Processing

Detection and Imaging of Corrosion Defects in Steel Structures Based on Ultrasonic Digital Image Processing

Research on Application of Support Vector Machine Method in Bridge Pier Impact Test

Non-contact Identification Method for Carbon Steel Corrosion Grade of Transmission Tower Based on Hyperspectral Technology

Study on Seismic Damage Prediction Model of Urban Bridges Group Based on PSO-SVM