Concrete Crack Pixel Classification Using an Encoder Decoder Based Deep Learning Architecture

Billah, Umme Hafsa; Tavakkoli, Alireza; La, Hung Manh

doi:10.1007/978-3-030-33720-9_46

Cited by 13 publications

(36 citation statements)

References 24 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The use of different deep-learning frameworks has gained considerable attention in recent works related to concrete crack detection [157,180,182,189,209]. A deep-learning-based SSD Inception V2 and SSD MobileNet models for concrete road damage detection was developed in another recent study [155,159,210].…”

Section: Surface-level Analysis: Concrete Crack Detectionmentioning

confidence: 99%

“…Another recent study utilized a U-net-based fully connected CNN model for concrete crack detection [188]. Some of the most recent studies have made use of different encoder-decoder-based deep-learning architectures to improve the existing limitations of crack detection systems using a pixel-wise classification of concrete images [182,189]. [213][214][215].…”

Section: Surface-level Analysis: Concrete Crack Detectionmentioning

confidence: 99%

“…There is a considerable amount of research effort devoted towards concrete crack detection for the surface-level SHM of civil infrastructures in general and bridges in particular [ 37 , 48 , 51 , 88 , 163 , 180 , 181 , 182 , 183 , 184 , 185 , 186 , 187 , 188 , 189 , 190 , 191 , 192 , 193 , 194 , 195 , 196 , 197 , 198 , 199 , 200 , 201 , 202 , 203 , 204 , 205 , 206 , 207 , 208 , 209 , 210 , 211 , 212 ]. Due to the concrete-based composition of the majority of civil infrastructures (e.g., dams, roads, buildings, sewers, bridges, tunnels), the techniques developed for concrete crack detection within a particular type of concrete structure can also be generalized towards other civil infrastructures.…”

Section: Algorithms For Data Analysismentioning

confidence: 99%

See 2 more Smart Citations

Review of Non-Destructive Civil Infrastructure Evaluation for Bridges: State-of-the-Art Robotic Platforms, Sensors and Algorithms

Ahmed

Gucunski

2020

Sensors

Self Cite

View full text Add to dashboard Cite

The non-destructive evaluation (NDE) of civil infrastructure has been an active area of research in recent decades. The traditional inspection of civil infrastructure mostly relies on visual inspection using human inspectors. To facilitate this process, different sensors for data collection and techniques for data analyses have been used to effectively carry out this task in an automated fashion. This review-based study will examine some of the recent developments in the field of autonomous robotic platforms for NDE and the structural health monitoring (SHM) of bridges. Some of the salient features of this review-based study will be discussed in the light of the existing surveys and reviews that have been published in the recent past, which will enable the clarification regarding the novelty of the present review-based study. The review methodology will be discussed in sufficient depth, which will provide insights regarding some of the primary aspects of the review methodology followed by this review-based study. In order to provide an in-depth examination of the state-of-the-art, the current research will examine the three major research streams. The first stream relates to technological robotic platforms developed for NDE of bridges. The second stream of literature examines myriad sensors used for the development of robotic platforms for the NDE of bridges. The third stream of literature highlights different algorithms for the surface- and sub-surface-level analysis of bridges that have been developed by studies in the past. A number of challenges towards the development of robotic platforms have also been discussed.

show abstract

Section: Surface-level Analysis: Concrete Crack Detectionmentioning

confidence: 99%

Section: Surface-level Analysis: Concrete Crack Detectionmentioning

confidence: 99%

Section: Algorithms For Data Analysismentioning

confidence: 99%

See 1 more Smart Citation

Review of Non-Destructive Civil Infrastructure Evaluation for Bridges: State-of-the-Art Robotic Platforms, Sensors and Algorithms

Ahmed

Gucunski

2020

Sensors

Self Cite

View full text Add to dashboard Cite

show abstract

“…The rare occurrence and aberrant shape of these defects demonstrate the difficulty in utilizing generic pattern recognition approaches for detection and segmentation. Therefore, accurately detecting these anomalous appearances with statistical estimation techniques is a challenging task [ 4 ]. On the other hand, considering crack detection as an anomaly detection problem allows for devising suitable approaches for accurately and efficiently localizing such defects.…”

Section: Introductionmentioning

confidence: 99%

“…Although using elaborate machine learning approaches helps with increasing the accuracy of defect detection, these techniques are computationally expensive and require appropriate parameter selection. In addition, these techniques require effective preprocessing of the input images, engineering suitable feature spaces capable of discriminating crack features from other normal pixel intensity differences or textures, and effective training [ 4 ].…”

Section: Introductionmentioning

confidence: 99%

Deep Learning-Based Feature Silencing for Accurate Concrete Crack Detection

Billah

Tavakkoli

2020

Sensors

Self Cite

View full text Add to dashboard Cite

An autonomous concrete crack inspection system is necessary for preventing hazardous incidents arising from deteriorated concrete surfaces. In this paper, we present a concrete crack detection framework to aid the process of automated inspection. The proposed approach employs a deep convolutional neural network architecture for crack segmentation, while addressing the effect of gradient vanishing problem. A feature silencing module is incorporated in the proposed framework, capable of eliminating non-discriminative feature maps from the network to improve performance. Experimental results support the benefit of incorporating feature silencing within a convolutional neural network architecture for improving the network’s robustness, sensitivity, and specificity. An added benefit of the proposed architecture is its ability to accommodate for the trade-off between specificity (positive class detection accuracy) and sensitivity (negative class detection accuracy) with respect to the target application. Furthermore, the proposed framework achieves a high precision rate and processing time than the state-of-the-art crack detection architectures.

show abstract

Deep Architecture Based Spalling Severity Detection System Using Encoder-Decoder Networks

Yasmin

2022

Advances in Visual Computing

View full text Add to dashboard Cite

Concrete Crack Pixel Classification Using an Encoder Decoder Based Deep Learning Architecture

Cited by 13 publications

References 24 publications

Review of Non-Destructive Civil Infrastructure Evaluation for Bridges: State-of-the-Art Robotic Platforms, Sensors and Algorithms

Review of Non-Destructive Civil Infrastructure Evaluation for Bridges: State-of-the-Art Robotic Platforms, Sensors and Algorithms

Deep Learning-Based Feature Silencing for Accurate Concrete Crack Detection

Deep Architecture Based Spalling Severity Detection System Using Encoder-Decoder Networks

Contact Info

Product

Resources

About