• Journal of the Society of Disaster Information
• /
• v.19 no.3
• /
• pp.498-509
• /
• 2023

Purpose: This paper proposes a method detecting the falling of a maintenance worker in the underground utility tunnel, by applying deep learning techniques using CCTV video, and evaluates the applicability of the proposed method to the worker monitoring of the utility tunnel. Method: Each rule was designed to detect the falling of a maintenance worker by using the inference results from pre-trained YOLOv5 and OpenPose models, respectively. The rules were then integrally applied to detect worker falls within the utility tunnel. Result: Although the worker presence and falling were detected by the proposed model, the inference results were dependent on both the distance between the worker and CCTV and the falling direction of the worker. Additionally, the falling detection system using YOLOv5 shows superior performance, due to its lower dependence on distance and fall direction, compared to the OpenPose-based. Consequently, results from the fall detection using the integrated dual deep learning model were dependent on the YOLOv5 detection performance. Conclusion: The proposed hybrid model shows detecting an abnormal worker in the utility tunnel but the improvement of the model was meaningless compared to the single model based YOLOv5 due to severe differences in detection performance between each deep learning model

• Fire Science and Engineering
• /
• v.30 no.2
• /
• pp.19-26
• /
• 2016

In this study, we conducted a low temperature operating test and miniature tunnel model test to study the fire detection capability and properties of an early fire detection system using an optical sensor linear detector that can be installed in harsh environments such as tunnel or utility-pipe conduits which are becoming the major and national infrastructure facilities. The test showed that the optical sensor linear detector was the only one functioned properly among five thermal detectors installed at a low temperature of $-20^{\circ}C$ for 5 days. To study were analyzed adaptability of optical sensor linear detector in the windy tunnel, the operating properties of the optical sensor linear detector when the wind velocity was varied between 0 m/s and 1 m/s in a miniature tunnel model. The temperature change was high when the wind velocity was 0 m/s.

• 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.

• The Journal of Engineering Geology
• /
• v.7 no.1
• /
• pp.27-36
• /
• 1997

This paper shows that crosshole seismic survey allows to detect even a small size of underground tunnel (about 2m$\times$2m). Such a small tunnel (e.g. infiltration tunnel) causes diffraction, as the seismic wave propagates, which results in distinctive variations of traveltime and amplitude of the first arrivals. This effect (or tunnel effect) is a typical indicator for the existence of tunnel and thereby an information about the tunnel location can be obtained. It was shown that the tunnel effect illustrated by numerical modeling (FDM) could be also observed in field measurements. The depth and shape of the tunnel were determined by a simplified processing method based on the use of amplitude variation of the first arrivals. The estimated location of the tunnel was well matched to that of the real tunnel.

• Proceedings of the KSR Conference
• /
• 2003.10b
• /
• pp.265-270
• /
• 2003

Though the probability of tunnel fire accident is very low, but critical fatalities are expected when it occurred. In this study the effect of critical safety parameters on tunnel fire accident are examined using probabilistic technique. Fire detection time, smoke spread velocity, passenger escape velocity, flash-over time, and emergency service arrival time are considered. In order to estimate the uncertainties of input parameters Monte Carlo simulation are used, and fatalities for each assumed accident scenarios are obtained as results. For the efficiency of iterative calculation PRA(Probabilistic Risk Analysis) code is developed in this study. As a result fire detection have large effect.

• Journal of the Korea institute for structural maintenance and inspection
• /
• v.2 no.4
• /
• pp.148-158
• /
• 1998

Determining the thickness if concrete lining and detecting of the cavity where is located behind tunnel lining plays an important role in the safety diagnosis of tunnel structure and the quality control. In this study, we made use of GPR and seismic method in order to find the cavity or flaw. Although GPR is very useful method in the concrete lining without rebar, it is difficult to detect the cavity in the reinforced concrete lining. We applied mini-seismic method to the reinforced concrete lining. The obtained seismic data was processed by means of seismic section in time domain and image section of power spectrum in frequency domain using Impact-Echo method as well. The proposed method can accurately show the location and depth of the cavity in the reinforced concrete lining.

• Geophysics and Geophysical Exploration
• /
• v.9 no.4
• /
• pp.279-290
• /
• 2006

The borehole radar methods used to tunnel detection are mainly classified into borehole radar reflection, directional antenna, crosshole scanning, and radar tomography methods. In this study, we have investigated the feasibility and limitation of each method to tunnel detection through case studies. In the borehole radar reflection data, there were much more clear diffraction signals of the upper wings than lower wings of the hyperbolas reflected from the tunnel, and their upper and lower wings were spreaded out to more than 10m higher and lower traces from the peaks of the hyperbolas. As the ratio of borehole diameter to antenna length increases, the ringing gets stronger on the data due to the increase in the impedance mismatching between antennas and water in the boreholes. It is also found that the reflection signals from the tunnel could be enhanced using the optimal offset distance between transmitter and receiver antennas. Nevertheless, the borehole radar reflection data could not provide directional information of the reflectors in the subsurface. Direction finding antenna system had a advantage to take a three dimensional location of a tunnel with only one borehole survey even though the cost is still very high and it required very high expertise. The data from crosshole scanning could be a good indicator for tunnel detection and it could give more reliable result when the borehole radar reflection survey is carried out together. The images of the subsurface also can be reconstructed using travel time tomography which could provide the physical property of the medium and would be effective for imaging the underground structure such as tunnels. Based on the results described above, we suggest a cost-effective field procedure for detection of a tunnel using borehole radar techniques; borehole radar reflection survey using dipole antenna can firstly be applied to pick up anomalous regions within the borehole, and crosshole scanning or reflection survey using directional antenna can then be applied only to the anomalous regions to detect the tunnel.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.8 no.8
• /
• pp.1179-1186
• /
• 2013

In this paper, an efficient vehicle detection and tracking algorithm for detection incident in tunnel is proposed. The proposed algorithm consists of three steps. The first one is a step for background estimates, low computational complexity and memory consumption Running Gaussian Average (RGA) is used. The second step is vehicle detection step, Adaboost algorithm is applied to this step. In order to reduce false detection from a relatively remote location of the vehicles, local features according to height of vehicles are used to detect vehicles. If the local features of an object are more than the threshold value, the object is classified as a vehicle. The last step is a vehicle tracking step, the Kalman filter is applied to track moving objects. Through computer simulations, the proposed algorithm was found that useful to detect and track vehicles in the tunnel.

• Journal of the Korean Society of Hazard Mitigation
• /
• v.11 no.1
• /
• pp.23-27
• /
• 2011

Due to the rapid development of construction technology with effective land utilization in this nation, many tunnels were and are being built across the country. However, the smoke and the heat generated from tunnel fire are the most important critical factors which may results in both massive personal injury and property damage, especially, due to the closed surrounding of the tunnel. Considering this particular nature of the tunnels, this study aims to install a fire detection system using an optic fiber cable to measure the temperature changes, compare, and analyze the resulted values with the times of temperature changes of the sensor by performing fire simulations under the same condition as a real fire test. From the results, it has been found that the temperature sensor detects a fire occurrence and generates an alarm within one minute after ignition for both a real fire test and a fire simulation alike, and also that the characteristics of temperature changes of the sensor has close relations with the speeds of the currents inside the tunnel. In addition, considering the tunnel fires can affect the evacuation efficiency and the fire extinguishing activities of the fire brigade inside the tunnel, the temperature sensor must be able to search and find the locations and directions of the fires correctly.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.7 no.5
• /
• pp.967-974
• /
• 2012

In this paper, a vehicle detection algorithm with HD CCTV camera images using GMM(Gaussian Mixture Model) algorithm and mathematical morphological processing is proposed. At the first stage, background could be estimated using GMM from CCTV input image signal and then object could be separated from difference image of the input image and background image. At the second stage, candidated object were reformed by using mathematical morphological processing. Finally, vehicle object could be detected using vehicle size informations depend on distance and vehicle type in tunnel. Through real experiments in tunnel, it is shown that the proposed system works well.