Crack Detection from a Concrete Surface Image Based on Semantic Segmentation Using Deep Learning

Yamane, Tatsuro; Chun, Pang‐jo

doi:10.3151/jact.18.493

Cited by 69 publications

(41 citation statements)

References 59 publications

(51 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…-The comparison of the detection results obtained in this article with the current work is presented in Table IV. The F1 score of the proposed model is found to be 25.8% and 5.26% higher than that of LightGBM and CrackUnet-19, 45.36% higher than that of the Semantic Segmentation based algorithm proposed by Yamane et al [45], and 20.06% higher than that of the Crack Segmentation Network proposed by Lee et al [46]. It also stands at the top while comparing with FCN and CNN-based algorithm given by the Islam et al [17] and Kim et al [47] respectively.…”

Section: Results Analysismentioning

confidence: 64%

Automatic Multiclass Instance Segmentation of Concrete Damage Using Deep Learning Model

2021

View full text Add to dashboard Cite

Concrete is one of the primary and most commonly used materials for the construction of buildings, roads, bridges, and dams. But it may lose its strength with age, moisture content, or due to other factors. As a result, little damage (crack and spall) to these structures may lead to sudden collapse or breakage, which in turn can decimate many lives with economic losses. Hence, to ensure their strength with the proper load, accurate screening of concrete surface damage is necessary for maintenance engineers to understand and evaluate the severity of the damage. Few preceding vision-based studies have proposed object detection and semantic segmentation approaches to carry out damage detection; however, the developed models could not segment different objects of the same class. The semantic segmentation approach separates regions that contain only objects of the same class. Concrete surface damage, on the other hand, may contain multiple objects, such as the Horizontal Crack, Vertical Crack, Diagonal Crack, Branch Crack, and Spall, that needs to be segmented separately so that specific measures can be taken. This study utilizes Mask Region-Based Convolutional Neural Network (Mask R-CNN) to manage concrete damage images and employ it to detect and segment defects in civil infrastructure with multiple objects. As we understand, this is the first research involving a Structural Health Monitoring (SHM) application for damage detection using instance segmentation. Our proposed model has been trained and tested on a dataset containing 800 images (480*480 pixels) of different types of crack and spall are collected from distinct structures of CSIR-CEERI, Pilani, campus. To test and validate the generosity, an additional 96 damage images are downloaded using Google. The pre-processed and annotated images are used to train and validate the Mask R-CNN Classifier. Our experimental results show that damage can be classified efficiently with 95.13% accuracy on a custom dataset and 96.87% on randomly picked images.

show abstract

Section: Results Analysismentioning

confidence: 64%

Automatic Multiclass Instance Segmentation of Concrete Damage Using Deep Learning Model

2021

View full text Add to dashboard Cite

show abstract

“…Chun et al [48] detected cracks in concrete surfaces using a light gradient boosting machine (LightGBM) considering pixel values and geometric shapes. You Only Look Once (YOLO), VGG Net, Inception Net, and Mask R-CNN have been frequently applied to detect concrete cracks in civil and infrastructure engineering studies [49].…”

Section: Cnnmentioning

confidence: 99%

CNN-Based Defect Inspection for Injection Molding Using Edge Computing and Industrial IoT Systems

Jeong

2021

Applied Sciences

View full text Add to dashboard Cite

Currently, the development of automated quality inspection is drawing attention as a major component of the smart factory. However, injection molding processes have not received much attention in this area of research because of product diversity, difficulty in obtaining uniform quality product images, and short cycle times. In this study, we proposed a defect inspection system for injection molding in edge intelligence. Using data augmentation, we solved the data shortage and imbalance problem of small and medium-sized enterprises (SMEs), introduced the actual smart factory method of the injection process, and measured the performance of the developed artificial intelligence model. The accuracy of the proposed model was more than 90%, proving that the system can be applied in the field.

show abstract

“…The developments might involve the latest scanning technologies [89,90] coupled with advanced techniques [artificial neural networks, deep learning, data fusion, etc.) for interpretation of the scan results [91][92][93][94][95][96][97][98]. In this context, the recent developments such as assessment of fire damage by changes in colour [99] and improvements in ultrasound techniques [100] might be considered.…”

Section: Assessment Of Fire-ravaged Concrete Structures With Non-destmentioning

confidence: 99%

Fire Resistance and Elevated Temperature in Reinforced Concrete Members: Research Needs for India

Dauji

Kulkarni

2021

J. Inst. Eng. India Ser. A

View full text Add to dashboard Cite

Fire and elevated temperature are important considerations for design of concrete structures, for both normal and nuclear facilities. In India, the present design guidelines for concrete subjected to fire and elevated temperature are limited to prescriptive minimum dimension and minimum cover. The other approach suggested is fire test based, and being resource intensive, is considered only for special cases of design. The various provisions of Indian standards are discussed and compared to the international standards. Fire test data for concrete/reinforcement properties at elevated temperature for Indian materials are scarce, and hence, the calculation methods are difficult to apply for Indian concrete. Validation of such methods for India is required, which again can only happen after generation of adequate database. The article highlights the present research needs for fire design of Indian concrete, along with development of analysis methodologies/guidelines for fire design of concrete structures, assessment of fire-affected concrete structures, and combination of the fire/temperature load with other extreme loads such as impact, seismic load, or missile load using advanced approaches. This article could serve as reference for designers as well as the starting point for several research initiatives in India and thereby is expected to improve the state of the art of fire design of concrete structures in India in the long term.

show abstract

Crack Detection from a Concrete Surface Image Based on Semantic Segmentation Using Deep Learning

Cited by 69 publications

References 59 publications

Automatic Multiclass Instance Segmentation of Concrete Damage Using Deep Learning Model

Automatic Multiclass Instance Segmentation of Concrete Damage Using Deep Learning Model

CNN-Based Defect Inspection for Injection Molding Using Edge Computing and Industrial IoT Systems

Fire Resistance and Elevated Temperature in Reinforced Concrete Members: Research Needs for India

Contact Info

Product

Resources

About