• Journal of Korean Tunnelling and Underground Space Association
• /
• v.19 no.1
• /
• pp.95-107
• /
• 2017

In this study, a preliminary study was undertaken for development of a tunnel incident automatic detection system based on a machine learning algorithm which is to detect a number of incidents taking place in tunnel in real time and also to be able to identify the type of incident. Two road sites where CCTVs are operating have been selected and a part of CCTV images are treated to produce sets of training data. The data sets are composed of position and time information of moving objects on CCTV screen which are extracted by initially detecting and tracking of incoming objects into CCTV screen by using a conventional image processing technique available in this study. And the data sets are matched with 6 categories of events such as lane change, stoping, etc which are also involved in the training data sets. The training data are learnt by a resilience neural network where two hidden layers are applied and 9 architectural models are set up for parametric studies, from which the architectural model, 300(first hidden layer)-150(second hidden layer) is found to be optimum in highest accuracy with respect to training data as well as testing data not used for training. From this study, it was shown that the highly variable and complex traffic and incident features could be well identified without any definition of feature regulation by using a concept of machine learning. In addition, detection capability and accuracy of the machine learning based system will be automatically enhanced as much as big data of CCTV images in tunnel becomes rich.

• Geophysics and Geophysical Exploration
• /
• v.14 no.2
• /
• pp.140-145
• /
• 2011

In order to analyze the influence of the noise on a cross-well traveltime tomography to detect a small scale low velocity body in a homogeneous medium, the first arrival travel times were computed one a tunnel model by a finite-difference ray tracing scheme. Three different types and four different intensity levels of white noises were added to the computed first arrival travel times, and velocity tomograms were constructed using an iterative inversion method (SIRT). Tomograms with the noise intensity up to 10% of the maximum traveltime delay in the tunnel model, showed the exact location of the tunnel. However, the velocity shown at the tunnel location was not close to air velocity but only slightly less than the velocity of the background medium. The additive random noise showed significantly less degree of influence on the resulting tomogram than the source- and receiver consistent noise.

• Journal of Korean Tunnelling and Underground Space Association
• /
• v.24 no.6
• /
• pp.583-598
• /
• 2022

As human-based tunnel inspections are affected by the subjective judgment of the inspector, making continuous history management difficult. There is a lot of deep learning-based automatic crack detection research recently. However, the large public crack datasets used in most studies differ significantly from those in tunnels. Also, additional work is required to build sophisticated crack labels in current tunnel evaluation. Therefore, we present a method to improve crack detection performance by inputting existing datasets into a deep learning model. We evaluate and compare the performance of deep learning models trained by combining existing tunnel datasets, high-quality tunnel datasets, and public crack datasets. As a result, DeepLabv3+ with Cross-Entropy loss function performed best when trained on both public datasets, patchwise classification, and oversampled tunnel datasets. In the future, we expect to contribute to establishing a plan to efficiently utilize the tunnel image acquisition system's data for deep learning model learning.

• Journal of Korean Tunnelling and Underground Space Association
• /
• v.22 no.6
• /
• pp.643-653
• /
• 2020

The concrete lining inspection by inspector after traffic control poses various problems such as congestion caused by traffic control and social loss costs resulting from it, and risks of traffic accidents and safety accidents for inspectors et al. To solve the problems, the concrete lining inspection has been carried out using automatic inspection equipment and image analysis that can be objectively and quantitatively investigated in overseas. In this study, to solve the problems of concrete lining inspection by inspector and to review improvement plan for inspection, inspection was carried out using automatic inspection equipment for ◯◯ tunnel that precision safety diagnosis has been conducted in 2019. Analysis was carried out for both inspection results. Automatic inspection equipment investigated defects in concrete lining such as cracks more accurately than precision safety diagnosis.

• Journal of Korean Society of Transportation
• /
• v.16 no.3
• /
• pp.15-24
• /
• 1998

도로망 전체구간의 소통 장애와 안전을 저해하는 요소 중의 하나로 터널구간을 들 수 있다. 특히 산악과 구릉지형이 많은 우리 나라의 도로설계에서 터널에 관한 교통특성이 중시되어야하지만 이에 대한 국내외의 기준이 명확히 정립되어 있지 못하다. 이러한 과제를 해결하기 위한 기초적인 시도로 본 논문에서는 영상처리기법에 의하여 양방 2차로와 4차로 의 도로에 포함된 터널부에 대하여 기본적인 교통특성을 조사·분석하였다. 양방 2차로에서 는 용량이 1,500대/시로 기본구간에 비하여 약 6%의 감소를 보이며, 4차로의 경우는 용량 이 2,000대/시로 기본구간에 비하여 10%의 감소를 보였다. 한편 교통밀도와 속도와의 관계 에 있어서는 양방 2차로의 경우 Underwood Edie의 모형이 그리고, 양방 4차로의 경우 Underwood의 모형이 비교적 설명력이 높은 것으로 나타났다.

• Journal of Korean Tunnelling and Underground Space Association
• /
• v.17 no.3
• /
• pp.393-401
• /
• 2015

In the construction of a TBM tunnel, it is very important to acquire accurate information of the excavated rock mass for an efficient and safe work. In this study, we developed the prediction system of TBM tunnel face ahead using probe drilling equipment and drilled hole imaging equipment to predict rock mass conditions of the tunnel face ahead. The prediction system consists of the probe drilling equipment, drilled hole imaging equipment and analysis software. The probe drilling equipment has been developed to be applicable to both non-coring and coring. Also the probe drilling equipment can obtain the drilling parameters such as feed pressure, torque pressure, rotation speed, drilling speed and so on. The drilling index is converted to the drilling index RMR through the correlation between a drilling index and core RMR. The developed system verification was carried out through a slope and tunnel field application. From the field application result, the non-coring is four times faster than a coring and the drilling index RMR and core RMR are similar in the distribution range. This system is expected to predict the rock mass conditions of the TBM tunnel face ahead very quickly and efficiently.

• Geophysics and Geophysical Exploration
• /
• v.18 no.1
• /
• pp.14-20
• /
• 2015

Three-dimensional (3D) resistivity method is an effective tool in the engineering site survey because it can provide a 3D resistivity distribution of the site. In this study, we tried to find out faults, fractures and coal seams that can cause the collapse of the tunnel. We carried out 2D resistivity survey along 5 parallel lines and 11 cross lines and merged all the apparent resistivity data for 3D inversion. Finally, from the 3D resistivity image and drilling data we presented the 3D distribution of faults, fractures and coal seams that are considered the main cause of the tunnel collapse.

• Tunnel and Underground Space
• /
• v.19 no.6
• /
• pp.517-525
• /
• 2009

Stereophotogrammetry is used to extract spatial information of an interested object by constructing a stereo-image from two or more photos. In this study, the stereophotogrammetry was adopted for a rock joint survey in mine tunnels. The orientations of discontinuities were measured from two mine tunnels with a clinocompass. To evaluate the effect of photographing light level on the stereophotogrammetry analysis, the light intensity was changed within a predefined range for every photograph. Those photographs were analyzed by using a commercial code for stereophotogrammetry - ShapeMetriX 3D, and the results from the analysis were compared with the manual measurement using a clinocompass.

• Journal of Korean Tunnelling and Underground Space Association
• /
• v.8 no.4
• /
• pp.377-388
• /
• 2006

Most structures experience deterioration after construction. A routine inspection and maintenance must be accomplished for the efficient use of the structures. The routine inspection will play a major role on the determination of maintenance period and method. This study aims development of an automated tunnel inspection system based upon a 3 dimensional laser scanner. As for the initial stage of the project, a prototype tunnel scanner has been developed. The development of a tunnel scanner prototype follows comparison between image scanning and laser scanning system and investigation on the applicability and adaptivity of the scanners to the railway tunnel scanner. The applicability of the laser scanner on the railway tunnel has been confirmed from the pilot test by using commercialized general purpose close range laser scanner and applicability of a laser scanner as a railway tunnel scanner has been checked. From the result, a prototype of railway tunnel scanner has been built and the calibration of the system was carried out. Finally the developed tunnel laser scanner has been applied to different shapes and sizes of tunnels in use.

• Journal of Korean Tunnelling and Underground Space Association
• /
• v.21 no.3
• /
• pp.419-432
• /
• 2019

Most of deep learning model training was proceeded by supervised learning, which is to train labeling data composed by inputs and corresponding outputs. Labeling data was directly generated manually, so labeling accuracy of data is relatively high. However, it requires heavy efforts in securing data because of cost and time. Additionally, the main goal of supervised learning is to improve detection performance for 'True Positive' data but not to reduce occurrence of 'False Positive' data. In this paper, the occurrence of unpredictable 'False Positive' appears by trained modes with labeling data and 'True Positive' data in monitoring of deep learning-based CCTV accident detection system, which is under operation at a tunnel monitoring center. Those types of 'False Positive' to 'fire' or 'person' objects were frequently taking place for lights of working vehicle, reflecting sunlight at tunnel entrance, long black feature which occurs to the part of lane or car, etc. To solve this problem, a deep learning model was developed by simultaneously training the 'False Positive' data generated in the field and the labeling data. As a result, in comparison with the model that was trained only by the existing labeling data, the re-inference performance with respect to the labeling data was improved. In addition, re-inference of the 'False Positive' data shows that the number of 'False Positive' for the persons were more reduced in case of training model including many 'False Positive' data. By training of the 'False Positive' data, the capability of field application of the deep learning model was improved automatically.