Assessment of Applicability of CNN Algorithm for Interpretation of Thermal Images Acquired in Superficial Defect Inspection Zones

Byeong-Su Jang; YoungSeok Kim; Sewon Kim; Hyun-Jun Choi; Hyung-Koo Yoon

doi:10.7843/kgs.2023.39.10.41

All Issue

2023 Vol.39, Issue 10 Preview Page Next Page

Assessment of Applicability of CNN Algorithm for Interpretation of Thermal Images Acquired in Superficial Defect Inspection Zones 포장층 이상구간에서 획득한 열화상 이미지 해석을 위한 CNN 알고리즘의 적용성 평가

31 October 2023. pp. 41-48

PDF XML

Abstract

The presence of abnormalities in the subgrade of roads poses safety risks to users and results in significant maintenance costs. In this study, we aimed to experimentally evaluate the temperature distributions in abnormal areas of subgrade materials using infrared cameras and analyze the data with machine learning techniques. The experimental site was configured as a cubic shape measuring 50 cm in width, length, and depth, with abnormal areas designated for water and air. Concrete blocks covered the upper part of the site to simulate the pavement layer. Temperature distribution was monitored over 23 h, from 4 PM to 3 PM the following day, resulting in image data and numerical temperature values extracted from the middle of the abnormal area. The temperature difference between the maximum and minimum values measured 34.8°C for water, 34.2°C for air, and 28.6°C for the original subgrade. To classify conditions in the measured images, we employed the image analysis method of a convolutional neural network (CNN), utilizing ResNet-101 and SqueezeNet networks. The classification accuracies of ResNet-101 for water, air, and the original subgrade were 70%, 50%, and 80%, respectively. SqueezeNet achieved classification accuracies of 60% for water, 30% for air, and 70% for the original subgrade. This study highlights the effectiveness of CNN algorithms in analyzing subgrade properties and predicting subsurface conditions.

Keywords

CNN

Infrared camera

Res Net 101

Squeeze Net

도로 하부에 발생된 이상구간은 사용자의 안전을 위협하고 보수하기 위해서도 많은 사회적 비용이 동반된다. 본 연구에서는 적외선 카메라를 사용하여 이상구간 매질에 따른 온도 분포를 실험적으로 평가하고 이를 머신러닝 기법으로 분석하고자 하였다. 대상 현장은 가로와 세로 및 깊이가 모두 50cm인 정육면체 형태로 설정하였고, 이상구간은 물과 공기로 결정하였다. 실험부지의 상부는 포장층을 모사하기 위해 콘크리트 블록을 설치하였으며, 오후 4시부터 다음날 오후 3시까지 총 23시간 동안 포장층의 온도 분포를 측정하였다. 측정된 값은 이미지 형태로 도출되었으며, 이미지 중간부분에서 측정 온도의 수치를 추출하였다. 최대온도와 최저온도의 차이는 물, 공기, 그리고 원 지반에서 각각 34.8°C, 34.2°C 그리고 28.6°C로 나타났으며, 이미지 분석 기법인 convolution neural network(CNN) 방법을 활용하여 각 측정 이미지에 해당하는 조건을 분류하였다. 분류를 수행하기 위해서는 res net 101과 squeeze net 네트워크가 이용되었다. res net 101의 분류 정확도는 물, 공기 그리고 원 지반에서 각각 70%, 50% 그리고 80%로 나타났고, squeeze net의 분류 정확도는 60%, 30% 그리고 70%로 나타났다. 해당 연구 결과는 수치데이터로 특징 판단이 어려울 경우 이미지 기반의 CNN 알고리즘을 활용하면 매질 특성 분석이 가능하고 지반내 상태도 예측할 수 있는 방법론을 보여준다.

References

Batayneh, A. T., Abueladas, A. A., and Moumani, K. A. (2002), Use of Ground-penetrating Radar for Assessment of Potential Sinkhole Conditions: an Example from Ghor al Haditha Area, Jordan, Environmental Geology, Vol.41, No.8, pp.977-983. 10.1007/s00254-001-0477-8

Bianco, S., Cadene, R., Celona, L., and Napoletano, P. (2018), Benchmark Analysis of Representative Deep Neural Network Architectures, IEEE access, Vol.6, pp.64270-64277. 10.1109/ACCESS.2018.2877890

Cheng, Q., Tang, C. S., Lin, Z. Z., Tian, B. G., and Shi, B. (2022), Measurement of Water Content at Bare Soil Surface with Infrared Thermal Imaging Technology, Journal of Hydrology, Vol.615, 128715. 10.1016/j.jhydrol.2022.128715

Cho, N. J., Cha, W., and Kim, H. K. (2016), Non-destructive Detection of Underground Cavities Using Thermal Images. Electron, Journal of the Korean Geotechnical Society, Vol.21, No.16, pp.5465-5476.

Frodella, W., Lazzeri, G., Moretti, S., Keizer, J., and Verheijen, F. G. (2020), Applying Infrared Thermography to Soil Surface Temperature Monitoring: Case Study of a High-resolution 48 h Survey in a Vineyard (Anadia, Portugal), Sensors, Vol.20, No.9, 2444. 10.3390/s2009244432344911PMC7250030

Go, G. H. and Lee, S. J. (2021), A Study on Numerical Analysis for GPR Signal Characterization of Tunnel Lining Cavities, Journal of the Korean Geotechnical Society, Vol.37, No.10, pp.65-76.

Hoai, N. V., Dung, N. M., and Ro, S. (2019, July), Sinkhole Detection by Deep Learning and Data Association. In 2019 Eleventh International Conference on Ubiquitous and Future Networks (ICUFN) (pp.211-213). IEEE. 10.1109/ICUFN.2019.8806136

Kang, S., Lee, J. S., Lee, S. J., Park, Y. K., and Hong, W. T. (2017), The Effect of Directivity of Antenna for the Evaluation of Abnormal Area Using Ground Penetrating Radar, Journal of the Korean Geotechnical Society, Vol.33, No.11, pp.1-4.

Kalhor, D., Ebrahimi, S., Tokime, R. B., Mamoudan, F. A., Bélanger, Y., Mercier, A., and Maldague, X. (2021), Cavity Detection in Steel-pipe Culverts Using Infrared Thermography, Applied Sciences, Vol.11, No.9, 4051. 10.3390/app11094051

Lee, E. J., Shin, S. Y., Ko, B. C., and Chang, C. (2016), Early Sinkhole Detection Using a Drone-based Thermal Camera and Image Processing, Infrared Physics & Technology, Vol.78, pp.223-232. 10.1016/j.infrared.2016.08.009

Lee, D. Y. and Cho, N. K. (2016), Understanding of Subsurface Cavity Mechanism due to the Deterioration of Buried Pipe, Journal of the Korean Geotechnical Society, Vol.32, No.12, pp.33-43. 10.7843/kgs.2016.32.12.33

Lin, R. (2022, February), Analysis on the selection of the appropriate batch size in CNN neural network. In 2022 International Conference on Machine Learning and Knowledge Engineering (MLKE) (pp.106-109). IEEE. 10.1109/MLKE55170.2022.00026

Information

Publisher :The Korean Geotechnical Society
Publisher(Ko) :한국지반공학회
Journal Title :Journal of the Korean Geotechnical Society
Journal Title(Ko) :한국지반공학회 논문집
Volume : 39
No :10
Pages :41-48
Received Date : 2023-08-29
Revised Date : 2023-09-17
Accepted Date : 2023-09-17
DOI :https://doi.org/10.7843/kgs.2023.39.10.41

Journal of the Korean Geotechnical Society ISSN:1229-2427(Print) 2288-646X(Online) 한국지반공학회 논문집

All Issue