Performance Comparison of Pretrained Convolutional Neural Networks on Crack Detection in Buildings

Date

2018-07-01

Author

Özgenel, Çağlar Fırat
Sorguç, Arzu

Metadata

Show full item record

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Item Usage Stats

385
views

0
downloads

Crack detection has vital importance for structural health monitoring and inspection of buildings. The task is challenging for computer vision methods as cracks have only low-level features for detection which are easily confused with background texture, foreign objects and/ or irregularities in construction. In addition, difficulties such as inhomogeneous illumination and irregularities in construction present an obstacle for fully autonomous crack detection in the course of building inspection and monitoring. Convolutional neural networks (CNN s) are promising frameworks for crack detection with high accuracy and precision. Furthermore, being able to adapt pretrained networks to custom tasks by means of transfer learning enables users to utilize CNN s without the requirement of deep understanding and knowledge of algorithms. Yet, acknowledging the limitations and points to consider in the course of employing CNN s have great importance especially in fields which the results have vital importance such as crack detection in buildings. Within the scope of this study, a multidimensional performance analysis of highly acknowledged pretrained networks with respect to the size of training dataset, depth of networks, number of epochs for training and expandability to other material types utilized in buildings is conducted. By this means, it is aimed to develop an insight for new researchers and highlight the points to consider while applying CNN s for crack detection task.

Subject Keywords

Transfer learning, Convolutional neural networks, Crack detection in buildings

URI

https://hdl.handle.net/11511/35867

DOI

https://doi.org/10.22260/isarc2018/0094

Collections

Department of Architecture, Conference / Seminar

Suggestions

OpenMETU
Core

Development and experimental verification of a structural health monitoring system for composite beams with embedded fibre bragg grating sensors Değerliyurt, Boray; Şahin, Melin; Department of Aerospace Engineering (2017) Structural Health Monitoring (SHM) is a discipline of development and application of monitoring and detecting adverse changes and damage in a structure in operation. This study explains development of a SHM system of composite beams with embedded Fibre Bragg Grating (FBG) sensors and its verification through experiments. There are considerations that require attention during manufacturing of composite specimens with embedded sensors. During manufacturing, protective teflon tubes are added to the ingress and...
Representation of damage information for post-earthquake damage assessment of reinforced concrete frames Anıl, Engin Burak; Akıncı, Burcu; Garett, James; Kurç, Özgür (2013-06-23) Current practice of post-earthquake damage assessment is based on visual inspections by experts. Research has shown that visual inspection and reports produced as a result of inspections can be subjective, incomplete, and vary in terms of the level of detail and thoroughness. Critical data that provides information about damage can be captured using laser scanners and processed for a more objective and information driven damage assessment practice. In addition to damage information, the damage severity iden...
Face Recognition Based on Embedding Learning Karaman, Kaan; Koc, Aykut; Alatan, Abdullah Aydın (2018-09-11) Face recognition is a key task of computer vision research that has been employed in various security and surveillance applications. Recently, the importance of this task has risen with the improvements in the quality of sensors of cameras, as well as with the increasing coverage of camera networks setup everywhere in the cities. Moreover, biometry-based technologies have been developed for the last three decades and have been available on many devices such as the mobile phones. The goal is to identify peop...
Characterization of laser scanners for detecting cracks for post-earthquake damage inspection Anil, Engin Burak; Akinci, Burcu; Garrett, James H.; Kurç, Özgür (2013-01-01) Objective, accurate, and fast assessment of damage to buildings after an earthquake is crucial for timely remediation of material losses and safety of occupants of buildings. Laser scanners are promising sensors for collecting geometrical data regarding the damaged states of buildings, as they are able to provide high coverage and accuracy at long ranges. Yet, we have limited knowledge on the performance of laser scanners for detecting earthquake damage, and requirements of such data collection. This paper ...
Deep learning-based reconstruction methods for near-field MIMO radar imaging Manisalı, İrfan; Öktem, Sevinç Figen; Department of Electrical and Electronics Engineering (2022-6-29) Near-field multiple-input multiple-output (MIMO) radar imaging systems are of interest in diverse fields such as medicine, through-wall imaging, airport security, and surveillance. These computational imaging systems reconstruct the three-dimensional scene reflectivity distribution from the radar data. Hence their imaging performance largely depends on the image reconstruction method. The analytical reconstruction methods suffer from either low image quality or high computational cost. In fact, sparsity-bas...

Citation Formats

Ç. F. Özgenel and A. Sorguç, “Performance Comparison of Pretrained Convolutional Neural Networks on Crack Detection in Buildings,” 2018, Accessed: 00, 2020. [Online]. Available: https://hdl.handle.net/11511/35867.