Civil Infrastructure Damage and Corrosion Detection: An Application of Machine Learning

Munawar, Hafiz Suliman; Ullah, Fahim; Shahzad, Danish; Heravi, Amirhossein; Qayyum, Siddra; Akram, Junaid

doi:10.3390/buildings12020156

Cited by 46 publications

(21 citation statements)

References 43 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Therefore, the F-score was used as the overall performance index for the developed model, which is defined by the harmonic mean of precision and recall, as expressed in Eq. 1 21) . According to this equation, the Fscores for each class were calculated.…”

Section: (A2)mentioning

confidence: 99%

Thickness Classifier on Steel in Heavy Melting Scrap by Deep-learning-based Image Analysis

Daigo

Murakami

Tajima³

et al. 2023

ISIJ Int.

View full text Add to dashboard Cite

Avoiding the contamination of tramp elements in steel requires the non-ferrous materials mixed in steel scrap to be identified. For this to be possible, the types of recovered steel scrap used in the finished product must be known. Since the thickness and diameter of steel are important sources of information for identifying the steel type, in this study, the aim is to employ an image analysis to detect the thickness or diameter of steel without taking measurements. A deep-learning-based image analysis technique based on a pyramid scene parsing network was used for semantic segmentation. It was found that the thickness or diameter of steel in heavy steel scrap could be effectively classified even in cases where the thickness or diameter of the cross-section of steel could not be observed. In the developed model, the best F-score was around 0.5 for three classes of thickness or diameter: less than 3 mm, 3 to 6 mm, and 6 mm or more. According to our results, the F-score for the class of less than 3 mm class was more than 0.9. The results suggest that the developed model relies mainly on the features of deformation. While the model does not require the crosssection of steel to predict the thickness, it does refer to the scale of images. This study reveals both the potential of image analysis techniques in developing a network model for steel scrap and the challenges associated with the procedures for image acquisition and annotation.

show abstract

Section: (A2)mentioning

confidence: 99%

Thickness Classifier on Steel in Heavy Melting Scrap by Deep-learning-based Image Analysis

Daigo

Murakami

Tajima³

et al. 2023

ISIJ Int.

View full text Add to dashboard Cite

show abstract

“…Where, 1 -5 denote five key construction scenes on the bridge. 1 denotes construction surveying and proofing, 2 denotes embedding steel casing, 3 denotes erect formwork and stand, 4 denotes steel bar welding and binding, and 5 denotes beam transportation and assembly.…”

Section: Formulation Of Identification Rules For Key Construction Sce...mentioning

confidence: 99%

“…Most current researchers use deep learning methods in artificial intelligence (AI) for safety monitoring during construction processes [4], with a focus on target detection of construction workers wearing helmets and holding equipment [5]. However, this method ignores the interrelationship between workers and construction objects, leading to a lack of early warning capability for safety monitoring when workers perform non-compliant construction operations.…”

Section: Introductionmentioning

confidence: 99%

Visual Relationship-Based Identification of Key Construction Scenes on Highway Bridges

Wang

Geng

et al. 2022

Buildings

View full text Add to dashboard Cite

Highway bridges play an important role in traffic construction; however, accidents caused by bridge construction occur frequently, resulting in significant loss of life and property. The identification of bridge construction scenes not only keeps track of the construction progress, but also enables real-time monitoring of the construction process and the timely detection of safety hazards. This paper proposes a deep learning method in artificial intelligence (AI) for identifying key construction scenes of highway bridges based on visual relationships. First, based on the analysis of bridge construction characteristics and construction process, five key construction scenes are selected. Then, by studying the underlying features of the five scenes, a construction scene identification feature information table is built, and construction scene identification rules are formulated. Afterward, a bridge key construction scene identification model (CSIN) is built; this model comprises target detection, visual relationship extraction, semantic conversion, scene information fusion, and identification results output. Finally, the effectiveness of the proposed method is verified experimentally. The results show that the proposed method can effectively identify key construction scenes for highway bridges with an accuracy rate of 94%, and enable the remote intelligent monitoring of highway bridge construction processes to ensure that projects are carried out safely.

show abstract

“…When combined, UAVs and CNNs can quickly detect, classify, and locate objects in images. (1)(2)(3) Owing to these advantages, various studies have been conducted in areas such as flood mapping, (4) agriculture, (5)(6)(7)(8)(9) the environment, (10) traffic, (11)(12)(13)(14)(15) structure management, (16)(17)(18)(19)(20) archaeology, (21) and disaster management (22,23) using drones and CNNs.…”

Section: Introductionmentioning

confidence: 99%

Positioning Errors of Objects Measured by Convolution Neural Network in Unmanned Aerial Vehicle Images

Kang¹,

Kim²,

Yun³

et al. 2022

Sensors and Materials

View full text Add to dashboard Cite

The conversion of unmanned aerial vehicles (UAVs, also called drones) and convolution neural network (CNN) facilitates the location of objects in real time using their sensors. In photogrammetry, the positional accuracy of objects is directly affected by the use of technology. It is necessary to improve the accuracy of object positioning to increase the utilization of drones and CNNs. In this study, the error factors that impede accuracy, such as the global navigation satellite system (GNSS) error of the UAV, camera distortion error, and camera posture error, were analyzed to improve the accuracy of object positioning. The effect of each error was also analyzed. The study was conducted in stages, such as establishing a method for the positioning of objects, specifying errors, and analyzing the amount and effect of error elements. The magnitude of the positioning errors was found by comparing it with accurate values measured by GNSS. Furthermore, the correlation of the errors with the factors that impeded accuracy was analyzed. Consequently, the effect of each error factor on the overall error was identified. These results can play an important role in improving positioning accuracy and developing UAV and CNN technologies employing sensors in the future.

show abstract

Civil Infrastructure Damage and Corrosion Detection: An Application of Machine Learning

Cited by 46 publications

References 43 publications

Thickness Classifier on Steel in Heavy Melting Scrap by Deep-learning-based Image Analysis

Thickness Classifier on Steel in Heavy Melting Scrap by Deep-learning-based Image Analysis

Visual Relationship-Based Identification of Key Construction Scenes on Highway Bridges

Positioning Errors of Objects Measured by Convolution Neural Network in Unmanned Aerial Vehicle Images

Contact Info

Product

Resources

About