Nothing Special   »   [go: up one dir, main page]
Next Article in Journal
Assessing Heterogeneity Among Cyclists Towards Importance of Bicycle Infrastructural Elements in Urban Areas
Previous Article in Journal
Public Involvement in Transportation Decision Making: A Comparison between Baghdad and Tehran
You seem to have javascript disabled. Please note that many of the page functionalities won't work as expected without javascript enabled.
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Infrastructures
Volume 9
Issue 9
10.3390/infrastructures9090152
infrastructures-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles thumb_up
...
Endorse textsms
...
Comment
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Application of PS-InSAR and Diagnostic Train Measurement Techniques for Monitoring Subsidence in High-Speed Railway in Konya, Türkiye

by
Gokhan Kizilirmak
1,2,* and
Ziyadin Cakir
3
1
Doctorate Program of Satellite Communication and Remote Sensing, Communication Systems Department, Graduate School, Istanbul Technical University (ITU), 34469 Istanbul, Türkiye
2
Turkish State Railways (TCDD), Railway Research and Technology Center (DATEM), 06050 Ankara, Türkiye
3
Department of Geological Engineering, Istanbul Technical University (ITU), 34469 Istanbul, Türkiye
*
Author to whom correspondence should be addressed.
Infrastructures 2024, 9(9), 152; https://doi.org/10.3390/infrastructures9090152
Submission received: 19 July 2024 / Revised: 25 August 2024 / Accepted: 3 September 2024 / Published: 7 September 2024
Browse Figures
Figure 1
<p>Photograph depicting a section of the Ankara–Konya High-Speed Railways provided by the Gokhan Kizilirmak.</p> "> Figure 2
<p>Geological maps: (<b>a</b>) shows the 1st study area; (<b>b</b>) shows the 2nd study area.</p> "> Figure 3
<p>Roger-800 performing measurements on the Ankara–Konya high-speed railway. The image was provided by Gokhan Kizilirmak.</p> "> Figure 4
<p>Photo showing the position of the laser measurement sensors. The image is provided by the Gokhan Kizilirmak.</p> "> Figure 5
<p>Illustration showing the working principle of levelling on a diagnostic train.</p> "> Figure 6
<p>Displacement diagram of the railway in the line of sight and at multiple passes.</p> "> Figure 7
<p>Simplified workflow of PS-InSAR processing in SARPROZ© (adapted from [<a href="#B58-infrastructures-09-00152" class="html-bibr">58</a>,<a href="#B59-infrastructures-09-00152" class="html-bibr">59</a>]).</p> "> Figure 8
<p>Image graphs for each time-series data stack: (<b>a</b>) CSK ascending; (<b>b</b>) S1–B/T65 descending; and (<b>c</b>) S1–B /T160 ascending. They show the 2D spatiotemporal baseline (yyyymmdd) spaces. Each point displays a scene, and each line displays an interferogram concerning a single master, which is represented with a red color dot.</p> "> Figure 9
<p>Reflectivity map showing the reference point location, city center and railway with blue colored text from all radar images.</p> "> Figure 10
<p>PSC maps and scatter plots: (<b>a</b>) CSK; (<b>b</b>) S1–B/T65; (<b>c</b>) S1–B/T160. PSC maps (red line means the railway) and mean velocity maps for CSK and S-1 analyses in LOS direction (dark blue line represents the 30 km-long railway).</p> "> Figure 11
<p>Vertical accumulated subsidence profiles of the railway along the 1st and 2nd study areas: (<b>a</b>) CSK; (<b>b</b>) S1–B/T65; (<b>c</b>) S1–B/T160; and (<b>d</b>) diagnostic train measurement time-series graphs.</p> "> Figure 12
<p>Accumulated subsidence graphs for the clustered PSs, where blue color represents PSs from CSK, orange color signs PSs from S1–B /T160, and lastly, grey color denotes PSs from S1–B /T65: (<b>a</b>) Location#1; (<b>b</b>) Location#2; (<b>c</b>) Location#3; and (<b>d</b>) Location#4.</p> "> Figure 13
<p>Specialized workflow model.</p> "> Figure A1
<p>Map of Ankara–Konya High-Speed railway showing the study areas.</p> ">
Versions Notes

Abstract

:
Large-scale man-made linear structures like high-speed railway lines have become increasingly important in modern life as a faster and more comfortable transportation option. Subsidence or longitudinal levelling deformation problems along these railway lines can prevent the line from operating effectively and, in some cases, require speed reduction, continuous maintenance or repairs. In this study, the longitudinal levelling deformation of the high-speed railway line passing through Konya province (Central Turkey) was analyzed for the first time using the Persistent Scatter Synthetic Aperture Radar Interferometry (PS-InSAR) technique in conjunction with diagnostic train measurements, and the correlation values between them were found. In order to monitor potential levelling deformation along the railway line, medium-resolution, free-of-charge C-band Sentinel-1 (S-1) data and high-resolution, but paid, X-band Cosmo-SkyMed (CSK) Synthetic Aperture Radar (SAR) data were analyzed from the diagnostic train and reports received from the relevant maintenance department. Comparison analyses of the results obtained from the diagnostic train and radar measurements were carried out for three regions with different deformation scenarios, selected from a 30 km railway line within the whole analysis area. PS-InSAR measurements indicated subsidence events of up to 40 mm/year along the railway through the alluvial sediments of the Konya basin, which showed good agreement with the diagnostic train. This indicates that the levelling deformation of the railway and its surroundings can be monitored efficiently, rapidly and cost-effectively using the InSAR technique.

1. Introduction

Railway subsidence monitoring has drawn a lot of attention from across the globe and has grown to be a significant issue due to the loss of both lives and financial investments in various places [1]. High-speed railways require very robust infrastructure to accommodate trains operating at speeds of up to 300 km/h, unlike conventional railways. Therefore, it is necessary to regularly monitor the geometric displacement of railways. Furthermore, labor expenses and the monitoring approach’s repeatability, accuracy, and efficiency should be taken into account. As an auxiliary monitoring approach, Interferometric Synthetic Aperture Radar (InSAR) methods, which are capable of measuring ground displacement at the millimeter level [2] and detecting track settlement or subsidence across whole railway networks [3,4], for monitoring transportation infrastructures have gained popularity in recent time owing to the open and free Sentinel-1 (S-1) data policy of the European Space Agency [5]. In [4], Chang et al. introduced a probabilistic method for post-processing InSAR time-series, designed to efficiently analyze data and detect structural instabilities in railway networks. This approach was validated across the entire railway network of the Netherlands, utilizing 213 Radarsat-2 acquisitions from 2010 to 2015. Over 3000 km of railway sections were analyzed, marking the first nationwide satellite-based application for railway monitoring.
One of the main benefits of the InSAR remote sensing technique is that measurements of targets can be made repeatedly over large regions more effectively than traditional geodetic terrestrial measurement techniques such as levelling and theodolite. Terrestrial levelling measurements in the field and Global Navigation Satellite Systems (GNSSs) are highly precise in estimating the rate of subsidence along transportation lines but time-consuming and expensive when considering hundreds of kilometers of railways. In addition, InSAR eliminates the need to close down the railway for traffic and enables the collection of a significant amount of information that is transmitted regularly and on schedule. In recent years, a great deal of research has been performed on ground displacement, with a focus on linear-shaped infrastructures like railways and highways. These studies aim to monitor a range of factors, such as the structural strength and deformation of the rail [4,5,6,7,8,9,10,11,12,13], assess the impact of tectonic and seismic movement on the rail [14], and determine the effect of ground subsidence caused by extraction of groundwater and human activity on the rail [15,16,17].
Radar interferometry has been utilized in the following ways: Small BAseline Subsets (SBAS) [14,17,18], Persistent Scatterer Synthetic Aperture Radar Interferometry (PS-InSAR) [3,6,7,8,9,12,19,20,21,22,23] and Differential Interferometric SAR (DInSAR) [24]. Long-term monitoring of the geological and geomorphological conditions surrounding the railway is important to ensure its safe and normal operation and to prevent fatal events and financial losses [10,12,25]. In some of these studies, geodetic methods such as ground levelling, GNSS and LiDAR were added to the SAR dataset(s), but no study comparing the InSAR method with the diagnostic train has been conducted so far. Therefore, the most important contribution and originality of this study is to include and use the diagnostic train in comparative analyses and to share the results. In addition, this work presents a unique study methodology by focusing only on railways, emphasizing the difference in the structure and the importance of observation precision.
The diagnostic train measures, along with InSAR, have been utilized to monitor certain parts of a high-speed railway line between the capital of Ankara and Konya in Türkiye. The diagnostic train data assessed in this study pertain to the Roger-800 model train, and it conducts regular geometric measurements of the railway lines. Additionally, this diagnostic train is capable of performing point deformation measurements using a laser beam source on the rails at intervals of 25 cm. The levelling measurements collected from the train are used to examine the engineering health state of the lines. This allows the engineers of railway administration to carry out maintenance or repair activities, as well as prepare for future actions. For these reasons, this study is based on data acquired from the diagnostic train as the reference measurement system. Although there is a focus on collecting frequent measurement data, it is important to note that the current method can only collect data from the line approximately once every three months. This means that it cannot perform measurements at a high frequency, observe its surroundings and detect rapidly developing subsidence events. Hence, it is advisable to contemplate employing an alternative monitoring technique that may offer more frequent and reliable data, regardless of weather conditions and without the need for personal field visits. Nowadays, this can be achieved by shortening the time period to hours instead of months with the available satellite constellation, depending on the location. InSAR measurements based on X-band Cosmo-SkyMed (CSK) and C-band S-1 satellites were compared with those obtained using the diagnostic train. PS-InSAR analyses were performed using the SARPROZ© version 2020 software, which is the abbreviation for the SAR PROcessing tool by periZ. Twenty (20) ascending direction CSK SAR images acquired from 15 December 2017 to 5 December 2019 were processed over Konya. Additionally, eighty-eight (88) images were retrieved in descending directions spanning from 11 December 2017 to 13 November 2020 for S1-B. In the last dataset, fifty-four (54) ascending mode S1-B SAR images, taken between 5 December 2017 and 7 December 2019, were analyzed.
Considering the previous studies conducted in the region using the InSAR technique, for example, Üstün et al. used GNSS locations and groundwater level variations to study the Konya region, and they found that GNSS displacements might reach −5 cm/yr [26]. While Comut et al. used the PS-InSAR technique and obtained displacement results ranging from −4 to −6 cm/yr in Konya’s city center, Üstün et al. used the DInSAR technique between 2002 and 2009, and they reported displacements of −3.4 cm/yr in Konya [27,28]. Using the SBAS technique, Caló et al. found a deformation value of −1.5 cm/yr in the Konya plain between 2002 and 2010. Sireci et al. reported land subsidence values reaching 11 cm/yr between 2014 and 2019 using Envisat-ALOS and S-1A/B images [29,30]. Moreover, Orhan found deformation values of up to −7.5 cm/yr in Konya between 2014 and 2018. In conclusion, according to the findings obtained from these studies for the Konya region, intense subsidence has been detected, and this situation has been associated with a rapid decrease in the groundwater level, as noted in some of these previous studies [31].
There are a total of three study areas within a 30 km length: a subsidence area, a stable region, determined using data from the diagnostic train and reports from the maintenance department between 2017 and 2019, and a third type, which lacks data from the train and is therefore analyzed only with the InSAR technique. The primary objective of this study is to identify a deformed region and a secondary region that remains stable, detected through the diagnostic train using the InSAR approach. Subsequently, the correlation between these two different techniques is examined based on the obtained results. The acquired strong correlation values highlight the benefits and advantages of utilizing InSAR as a reliable secondary measurement for monitoring subsidence in railways. Furthermore, based on the knowledge acquired in this study, a flow chart outlining the application phases of the InSAR technology, specifically adapted to the unique construction, maintenance, and repair needs of railways, is also provided.
This paper is organized as follows. First, the study area, its brief geological characteristics, and the measurement techniques are described in Section 2. Then, the InSAR results from both C-band and X-band datasets are reported in Section 3. In Section 4, the results, comparisons between the datasets, and their correlations are explained. Following this section, discussion and suggestions are presented in Section 5, along with an InSAR analysis evaluation proposal developed specifically for deformation monitoring on railways. Finally, a conclusion is given in Section 6.

2. Materials and Methods

2.1. Study Area and Its Geological Background

The study area is situated in Konya, in central Anatolia. The high-speed rail line spans 218 km and is designed for a maximum operating speed of 300 km/h or 190 mph, commencing its service on 23 August 2011 [32]. The study area with the datasets consisting of radar images is approximately 800 km2, and the length of the line passing through this area is 30 km. A section of the high-speed railway line connecting Ankara and Konya is shown in Figure 1.
As seen from Figure 1 and the map in Appendix A, most of this railway line passes through rural/agricultural areas. One of the most important problems in the region is the continuously increasing number of sinkholes. Karstification has caused sinkhole evolution to be recognized as a dynamic and natural process in Konya [33]. Additionally, groundwater has been decreasing as a result of artesian wells drilled inside the city. As an outcome, the area exhibits a linear deformation [34] and a fine-grained sediment soil structure, particularly in regions with dense agricultural land near the Konya plain. This phenomenon results in sinkholes and land subsidence because the sediments settle in response to groundwater withdrawal that occurs quickly [35]. Sinkhole is a detailed research subject, and since it is not the main research subject of this article, it is only briefly described here in order to get to know the region in more detail but will not go into more depth.
Through the analysis of measurements from the diagnostic train and reports obtained from the maintenance department, three distinct areas were identified along a 30 km extension of the track. The first area, referred to as the unstable area, exhibited instability. The second region, known as the stable area, showed no signs of instability. The third location, a station area, could not be evaluated due to the absence of data from the diagnostic train, but it was analyzed using the InSAR technique. The first section spans approximately 6 km and is situated between the rural regions of Asagipinarbasi and Yukaripinarbasi. The second test area is situated in the Selcuklu district of Konya, and it encompasses a 3.5 km length of the railway line that runs through the inner city. This part is primarily characterized by residential and industrial areas. Finally, the third analyzed segment pertains to the central train station in Konya City Centre (Appendix A).
An analysis of the geological conditions in the region was conducted in order to provide a comprehensive understanding of the subsidence situation. Geological maps show that the railway line passes through alluvial sediments, made up of mainly clay and silt, interbedded with sandstone and gravel in the Konya Basin (Figure 2a,b). Such clay and silt units are characterized by high compressibility, playing as a predisposing factor in the processes of subsidence induced by groundwater extraction. Figure 2a represents the geological structure of the first study area as ‘Qa’, and Figure 2b represents the geological structure of the second study area as ‘Qa’ and ‘Qae’. In the Turkish State of Railways report, the geological structure of the third study area, which is the range of Km: 181 + 190–212 + 555, is mostly covered by old alluvium ‘Qae’. Since the location of the relevant region corresponds to approximately km: 210+000, the definition of this km interval also includes the third study area. Moreover, old alluvium ‘Qae’ is generally composed of gravelly sand, clay and silt. The report also states that setting the railway on a weak substructure represented by the clayey levels in the alluvium should be considered as a potential subsidence problem. These clay units are characterized by their great compressibility, which makes them important factors in the consolidation and subsidence processes caused by groundwater extraction [36].
The lithology of the Konya Basin is mostly composed of fine-grained sediments, particularly clay, sands and silts, as explained and shown in the maps in Figure 2a,b in detail. According to the definition in the International Union of Railways (UIC) Code 719R, if the soil is clay and clay shale, the structure is represented in the QS1 quality class, and weak ground expression is used for this type of class [37] (Table 1). Therefore, it can be concluded that more budget and time resources should be spent on the repair and/or maintenance of the railway line in order to keep it stable in this region, which has a clayey and soft structure, i.e., weak and highly water permeable ground.

2.2. Dataset

2.2.1. Diagnostic Train Measurement

A diagnostic train that makes periodic measurements in the Directorate General of Turkish State of Railways (TCDD) is the measurement technique currently used to detect movement or irregularities in the rails. The train drives along one track and sends laser pulses to collect values in a highly precise way. The model of the train is called the Roger-800 of the Mermec Group, which is shown in Figure 3. The primary method for assessing track geometry in the majority of nations is through the use of diagnostic trains equipped with inertial measurement systems. Ankara–Konya high-speed railways in Türkiye have been monitored for their condition using the Roger-800 train since August 2017.
The diagnostic train operates in both directions with a maximum speed of 120 km/h. Basic technical information about the Roger-800 is given in Table 2 [38]. The geometry measuring system is placed on the front side of the car bodies, positioned between axles, as seen in Figure 4. Track geometry is measured at every 0.25 m and stored in onboard computer systems, where it is processed immediately during the ride. All measured data can also be plotted as a graphical representation and exported to external data media. The precision of the geometric measurements is 0.5 mm. One of the biggest advantages of the diagnostic train is that it can measure multiple parameters at the same time, including track geometry, gauge measurement, the condition of ballast, track videos/photos and the acceleration of bogies/car bodies with high precision. However, the diagnostic train can only observe the railway line itself, not its surroundings, and the temporal measurement frequency is relatively low, e.g., three or four times for high-speed lines and twice a year for regional lines. In addition to these limiting factors, the sensors can be adversely affected by extreme weather conditions, which can lead to noises in the measurements and, thus, misinterpretation of results. It should also be noted that there are also operating and maintenance costs, which include periodic inspections and calibrations.
This research includes a total of nine train measurements, taken on average every three months between December 2017 and December 2019.
The working principle of the levelling measurement sensors in the diagnostic train is similar to the principle of terrestrial levelling measurement, and here, the Inertial Measurement Unit (IMU) is used, as shown in Figure 5.

2.2.2. Multi-Temporal PS-InSAR

SAR is an active sensor, and the sensor emits a signal, which the surface reflects, and the same sensor then detects again, as shown in Figure 6. Multi-Temporal InSAR (MT-InSAR), an advanced InSAR approach, measures displacement more accurately by examining the phase information from a stack of SAR data collected over the same region [39]. It can reduce errors in deformation estimations and address some of the limitations of conventional InSAR, such as uncorrelated phase disturbances [40].
In SAR interferometry, the target region is scanned using a constellation of satellites with identical instrumental and orbital properties or through a single satellite to obtain sequential acquisitions over time. The frequency of measurements is thus a function of the frequency of satellite passes. The electromagnetic spectrum’s microwave region is where SAR satellites transmit their signals. The detected signal’s phase and amplitude are represented in its data. Interferometry is based on the creation of interferograms and uses the phase differences between acquisitions, i.e., at various times, to estimate the movement of the ground surface, as seen in Figure 6 [3,41,42].
As a result, deformations are found along the satellite’s line of sight (LOS). The purpose of InSAR processing is to calculate the real displacement component, ΦDispl, which is distinguished from other signal contributions as follows:
ΦDispl = ΦInt – ΦRTE – ΦAtm – ΦOrb – ΦNoise – 2 kπ,
where ΦInt denotes the wrapped interferogram, ΦRTE represents the residual topographic error component, ΦAtm signifies the atmospheric component of each image, and ΦOrb denotes the orbital error component. Phase ambiguity is resolved through phase unwrapping techniques since ΦNoise accounts for phase noise and 2 kπ regulates it [42]. To address these errors, noise and uncertainties encountered when trying to calculate the real displacement component, some auxiliary data and filters, which have been applied in SARPROZ©, such as the Digital Elevation Model (DEM), precise orbit information from the image acquisition time, and, if available, weather information including precipitation and humidity information, are used. For example, DEM data are used to analyze the topographic effect. For this analysis, since Konya is predominantly a flat region, the 30 m-SRTM DEM (Shuttle Radar Topography Mission Digital Elevation Model) data are sufficient to eliminate the topographic error. In order to eliminate the orbital effect, sensitive orbit data are used, and finally, the atmospheric contribution can be estimated and eliminated by analyzing a sequence of images obtained at different time intervals [2,42]. After eliminating the influences of topographic, atmospheric and residual noise, the remaining anomalies in phase can be attributed to subsidence occurring along the radar’s LOS, as shown in Figure 6. It should be noted that multi-pass interferometry methods, such as PS-InSAR, allow for the identification and elimination of these artifacts [43].
PSs have consistent scattering properties throughout time and a dominant reflection inside each pixel cell. This reduces both temporal and geometric decorrelation. Furthermore, the signal interference can be assessed and eliminated utilizing a sequence of images obtained at various time intervals [44]. Utilizing a series of acquires, the accuracy of detecting dependent pixels can also be enhanced [45,46].
The different InSAR techniques that are currently available can be broadly categorized as follows: (a) PS Interferometry, first developed by Ferretti et al. [40]; (b) Small—BAseline Subsets (SBASs) and its variants by Berardino et al. [47]; and (c) hybrids of PS-InSAR and SBAS, such as StaMPS (Stanford Method for Persistent Scatterers) [48] and SqueeSAR [49]. PS-InSAR processing aims to identify strong and consistent signals emanating from prominent reflectors, such as man-made structures, sharp objects and buildings. There are other evaluations of different strategies in the literature that go into more detail, e.g., [39,50]. In this paper, a standard PS-InSAR approach is proposed, focusing on the vertical deformation of the railway and the strong, persistent reflective behavior of the rails due to their metallic structure [1,9]. This strong reflectivity characteristic can be used to generate sufficient PSs for APS estimation and the PS-InSAR approach [34].
During these PS-InSAR analyses, SARPROZ© software was used, and the software mainly developed in MATLAB® R2015a can select different sets of interferograms to combine and process long data series and offers many different options and techniques [51]. Parameters can be estimated with the traditional PS-InSAR algorithm. All details about this interferometric package can be found in [52,53] and on the official website [54].

3. Data Processing of PS-InSAR

In the data processing, as shown in Figure 7, the first step involves converting datasets into a format readable using SARPROZ© software. Various tasks such as calculating dataset statistics, incorporating precise orbit data if available, automatically downloading and inserting weather data for each SAR scene into the analysis, and selecting a subset of images to cover the AOI are performed in these very initial steps. Within the co-registration process, offsets between the master and slave images are estimated by investigating the correlation of amplitude information in the spatial and spectral domains. In the next stage, each slave image is aligned to the master so that the portion of imaged terrain for each pixel in the slave image corresponds to the same pixels in the master image in PS-InSAR. The selection of the master image is conducted to ensure optimal InSAR coherence across the entire stack and within the barycentre of the distribution of normal/temporal baselines. From the co-registered stack of images, the reflectivity map, as the temporal average of the intensities of all images in the dataset, is generated, and the Amplitude Stability Index (ASI) is computed. Preliminary geocoding using 30 m-SRTM DEM data is then performed. Very precise geolocation of each PS point could be obtained due to proprietary procedures implemented in SARPROZ©. Next, the DEM in SAR coordinates is calculated to assist in PS-InSAR processing [1,22,45,46,51,52,53,54,55,56,57,58,59].
Twenty ascending-direction CSK SAR images between 15 December 2017 and 5 December 2019 were analyzed. Additionally, 88 images were retrieved in descending and ascending directions, spanning from 11 December 2017 to 13 November 2020, for S1-B/T65 (Track 65), as represented in Figure 8. However, the first two-year period of this S1-B/T65 analysis was considered to capture the period of the other datasets and the diagnostic train in this study. In the last dataset, 54 ascending-direction S1-B/T160 (Track 160) SAR images between 5 December 2017 and 7 December 2019 were analyzed. The temporal differences between datasets were kept quite low so that the differences in terms of deformation values would not increase too much (Table 3).
In detail, a total of 20 images from CSK obtained were tried to be spread over a period of two years, taking into account the archive status and the budget availability. Analyses were conducted over an area of around 800 km2 inside the province of Konya to examine the deformation of a 30 km railway line and its surroundings using three different radar datasets covering a similar region, as seen in Figure 9. The research region was restricted to 800 km2 mostly due to the limited coverage offered by the CSK Strip Map HIMAGE mode data, which has a scene size of 40 km, compared to the S-1 Interferometric Wide mode data, which has a swath width of 250 km.
In this work, a standard PS-InSAR approach [40] was adopted to estimate and remove the APS from every image’s phase signal. Briefly, the APS was fitted to the outputs of the displacement velocity model. For each PS, the LOS velocity, displacement time series and height were computed relative to a reference point identified in a more stable area outside the basin/Konya city centre, as shown in Figure 9. After conducting experimental analyses and reviewing previous studies in the region [26,27,28,29,30,31], it was determined that the reference point should be located outside Konya, particularly in mountainous areas. This will minimize the impact of subsidence displacements in and around the city center, as seen in Figure 9 and Figure 10. Furthermore, this reference point selection process involved evaluating sites with better temporal coherence values in comparison to others. The reflectivity map image represents the topography of the study area and contains the amplitude and phase information for each pixel of the SLC image [49]. Looking at the reflectivity map taken from CSK (X-band) as a sample, the railway, which is a linear-shaped and high-reflective target, is clearly visible in Figure 9.
To compensate for the effect of different incidence angles, the LOS mean deformation velocity was obtained and converted to the vertical direction using Equation (2). Based on train measurements, and previous studies [27,29,30,34,60,61] mentioned in Section 1, only the component of displacement should be evaluated as vertical displacement, ignoring horizontal displacement and any potential errors. In addition, LOS deformation is assumed to be purely vertical since the Konya Basin is quite flat and the ascending and descending Sentinel LOS velocities are almost identical, as shown on the right part of Figure 10. In another study conducted in the same region, the following statement is made for vertical transformation. If the deformation is unidirectional, LOS directional deformations that are generated as output may be simply transformed into actual deformation values [34].
V e r t i c a l   v a l u e = L O S   v a l u e c o s θ ,
where θ denotes the local incidence angle.
In the same whole study area, almost 74,000 and 99,000 PS points (PSs) could be properly estimated from the S1-B/T160 and also S1-B/T65 radar satellites, respectively, while almost 316,000 PSs were from the CSK dataset (Figure 10). The InSAR results obtained from the different datasets were compared to assess their internal consistency, and all results were checked and found to be in good agreement, as seen in the scatter plots in each right part of Figure 10a–c. The difference between the numbers of PSs taken from the different datasets is mainly caused by the higher spatial resolution of the CSK X-band data. Looking at the PSC density maps on the left of Figure 10a–c, it can be seen that the CSK dataset produces more PSs, but there are also many unstable PSs in agricultural and mountainous areas in the region, mainly due to the X-band characteristics.
In this analysis, only the PSs on the railway in the study regions were used, and since the temporal coherence of the railway is always stable due to its permanent metallic reflective structure, unstable PSs in areas other than the railway were ignored. Based on the measured amplitude, PSCs are chosen at the start of the processing. After that, their displacement is calculated together with the effects of noise, atmospheric contribution and height correction. PSCs are usually chosen from the candidate pool according to their degree of noise and coherence [40,48]. APS estimation plays a crucial role in PS-InSAR applications by facilitating the calculation of linear deformation velocities and compensating for topographic height effects [62]. For accurate APS estimation, it is advisable to use an appropriate threshold, typically ASI > 0.75, when selecting initial PSs [63]. Based on this information and for creating a network of PSCs, an ASI threshold value > 0.75 was used and applied to estimate preliminary parameters and APS. A larger set of points, based on the reflectivity map and spatial coherence, was used during the compensation of APS.
After APS removal, the final estimates of height and velocity were computed. At the step of visualizing the final results, a threshold ASI index > 0.60 on temporal coherence was used to include more PSs [64] (Figure 10). As a check stage, the reliability of the deformation estimations can be determined utilizing many independent satellite tracks that observe the same deformation signal, which is apparent at each right part of Figure 10 [65].
The first research location is a rural region, which results in decreased coherence with the X-band. However, it was noted that more consistent findings could be achieved from the C-band for this type of region, as shown in Figure 10. In the second region, the presence of residential and industrial structures leads to a greater number of reflecting objects compared to the predominantly rural first area. This situation led to the acquisition of more PSs located along the railway track. Upon further examination of the PSs in the third study region, it is evident that the coherence values are high, and the mean standard deviations are within an acceptable range. The analysis of the deformation maps in Figure 10 reveals that the first study area is indicated by a red circle, the second study area is depicted within a green rectangular box, and the third study area is situated near the Konya city center, denoted by an asterisk at the end of the line. However, there is a spot, located between 37.98° and 38° latitude and 32.6° longitude, detected through SAR analysis where no evidence of deformation has been found using the diagnostic train and the relevant maintenance department. This location is in close proximity to the railway and has been recognized as a problematic area due to the presence of deformation. This area is situated to the south of the line and mostly encompasses the Konya Organised Industrial Zone, as shown by a black arrow in Figure 10. This event once again highlights the significance and thoroughness of the deformation maps derived from InSAR analysis. The data acquired from the InSAR analysis will enable us to continuously monitor whether or not the scenario occurring around the line will have a direct impact on it. In summary, the second and third research regions yield more favorable outcomes, including improved coherence, increased number of PSs on the railway tracks and reduced standard deviation, compared to the first area, despite the findings being deemed satisfactory (Table 4).

4. Results

This analysis assumes the diagnostic train levelling measurement data as a reference, disregarding any potential flaws that might impact the data. Based on the European Norm [66], the stability of a rail segment for high-speed trains traveling at speeds of 220–300 km/h is determined. According to these criteria, a rail segment is considered stable if it maintains a structurally healthy condition with a settlement of less than 6 mm. Conversely, a rail segment is deemed potentially unstable if it has a settlement greater than 6 mm [5,66]. By comparing the results of the PS-InSAR analyses of the 30 km railway line and the diagnostic train measurements in areas 1 and 2, it is evident that areas 1 and 3 are not stable, whereas area 2 remains stable based on the cumulative results of the two-year time series graphs (Figure 11).
The correlation was assessed using the Pearson Product Moment Correlation Coefficient, and the determination of correlation relations between each dataset relies on the overlapping zones generated by the PSs in the profiles of the first and second research regions. The correlation coefficients for study area-1 between the diagnostic train and CSK analysis is 0.78, with S1-B/T65 at 0.93, and lastly with S1-B/T160 at 0.69. Hence, there exists a robust and favorable association between the datasets. For study area 2, the correlation coefficient between the diagnostic train and the CSK analysis is 0.88, with S1-B/T65 at 0.65, and lastly with S1-B/T160 at 0.87. Therefore, there is a strong and positive correlation between the datasets for the second study area (Table 5).
Since there is no measurement from the train for the third study area, the correlation coefficients were calculated only between the clustered PSs. This information is shown in Figure 12 and Table 6. For this correlation analysis based only on InSAR outputs, due to differences in measurement time, values one month apart are used as a basis by calculating the average deformation for the relevant PS. These PSs are specially chosen from railways that run through the station and are also utilized as parking lots. As seen in Table 6, there is a significant and robust positive correlation between these PSs obtained from different SAR datasets.

5. Discussion and Suggestion

A great performance of the advanced PS-InSAR remote sensing technique for monitoring the railway’s subsidence deformation has been shown, which had not been conducted before in the area of interest. By considering the results, C-band data, with its free access opportunity, would be a great choice for evaluating hundreds of kilometers of railway lines. However, for obtaining and focusing detailed information about displacements and representing it with more PSs, X-band data seem more suitable. In [67], the researchers mention that the increased PSI measurements’ reliability in relation to the infrastructures is a result of the PSs’ higher density. While the density of PSs on the railway line is sensitive to the orientation of the railways, observations from ascending and descending orbits would be helpful [21,22], as performed in this study. It can be seen from Figure 11 that in the second area, which is the right half parts of each graph, the subsidence values are close to or around zero, while the behavior of the PSs in this stable area is spatially homogeneous, representing minimal variations in the deformation rates.
As a suggestion, the complicated structure of SAR image types and characteristics necessitates careful consideration when assessing appropriate data sources for monitoring railway subsidence. In addition to data selection, it is crucial to perform image processing meticulously to ensure precise computation of deformation and deformation velocities. Phase residual plots, correlation graphs, elevation charts, error diagrams, standard deviation values, etc., should be checked during the analysis to ensure that phase residual errors and atmospheric effects are eliminated or minimized in the process, and if necessary, parameters and thresholds used should be changed, and after that, the analysis should be repeated. It should be noted that this is a relative technique, i.e., it is based on the behavior of the reference point, and the process goes on according to the filter values and parameters used. In addition, the whole structure of the study region, the number of images used, and the band characteristics also play an effective role in the analyses. Although the high correlation between the results of SAR datasets is important for validation, it is unrealistic to expect exactly the same result. The reason for this is that the X-band and C-band characteristics are different, which leads to the formation of different numbers of PSs by giving different responses to deformation analyses over the region. This causes a slight difference in the averages and accuracy of the results. Moreover, as seen in Figure 8, it should be noted that factors such as differences in the start and end dates of the datasets within the time series for a few days, the total number of images used, and different orbit baseline intervals will also cause minor differences in the outputs of the X-band and C-band datasets.
When observed in rural areas, the chosen X-band’s temporal consistency decreases, making it difficult to acquire consistent subsidence monitoring findings. This is due to the changing types and lengths of agricultural crops in this area over time. This study utilized an X-band CSK dataset comprising 20 SAR data collected over a span of two years. However, due to the long base distance in the repetitive passes of the satellite set and the temporal differences of two or three months between images (Figure 8), the desired efficiency could not be achieved. While a large amount of images does not necessarily guarantee more precise results, research explains a correlation between the number of images and the standard deviation as follows. In the research [65] conducted in Groningen/Netherlands, the standard deviation is ~0.1 mm/year per √km when 24 interferograms are used. However, in a subsequent analysis using 74 interferograms in the same location, the standard deviation decreased to ~0.04 mm/year per √km. Hence, adjusting the amount of data when dealing with the X-band dataset might be presented as a guideline in this study. Moreover, the lack of data collected from the descending orbit for the CSK dataset might be considered a limiting factor in this research. Similarly, the lack of diagnostic train data for the third study region is viewed as a drawback in comparative analyses. However, the absence of train data in regions like these has once again emphasized the crucial significance of InSAR remote sensing for monitoring structural deformation. While the subsidence phenomena in the whole study area are effectively demonstrated by our results across the datasets, they should still be improved via in situ analysis using techniques like GNSS and levelling as independent ground-truth data, which can be seen as another limitation of this study. Furthermore, due to the fact that railway lines typically pass both urban and rural regions, it is necessary to develop specific methods for monitoring deformation events in rural areas. This is where the existence of C-band observations, which can be obtained without any cost, becomes significant [13]. Artificial reflectors, such as corner reflectors, can be placed in the areas lacking PSs but which need to be monitored for deformation [68]. Corner reflectors offer a cost-effective alternative to on-site and ground-based monitoring equipment. They can be strategically installed in key sections of the structure to generate effective reflections, facilitating more efficient identification of these targets as PSs during the processing phases [69].
The analysis results are crucial for properly placing GNSS receivers at certain locations identified as critical, such as the first and third places mentioned in this study. Performing further geodetic observations in these crucial locations will enhance the level of detail and quality of the observations, enabling more precise decision-making. As an additional suggestion, if a more detailed geophysical study is desired for such areas, a Ground Penetrating Radar (GPR) device can be easily mounted on the diagnostic train. GPR is a well-recognized Non-Destructive Testing (NDT) method that causes minimum disruption to local traffic and is very efficient in examining the internal structure of the ballast layer, subgrade and embankment [3]. In [69], researchers review recent advancements and applications in satellite remote sensing and ground-based NDT methods for monitoring transport infrastructures in detail. This proposal appears to be a feasible approach for the Asagipinarbasi–Yukaripinarbasi region, which is referred to as the first research area in this study and has been identified as a critical area. Another important limitation of any InSAR technique is that tunnels cannot be measured due to SAR’s nature. Such structures will continue to be precisely and regularly monitored using diagnostic trains and/or terrestrial measurement techniques and various sensors.
The Workflow Chart is designed specifically for addressing railway-related issues and providing specialized solutions, as shown in Figure 13. It is anticipated that this chart will offer a straightforward and practical method for analysts or researchers who need to monitor subsidence events in and around railway areas utilizing an advanced InSAR technique as a complementary surveying and monitoring tool.

6. Conclusions

This study highlights the essential importance of the InSAR remote sensing approach for monitoring subsidence in railways and their surrounding environment, thanks to its very accurate millimeter precision and independence from weather conditions. In order to demonstrate the efficiency of the InSAR approach, three separate time-series datasets from S-1 and CSK radar satellites were utilized across three specific places along a 30 km length of the railway section. The vertical velocities derived from both S-1 and CSK data analysis demonstrate a strong internal agreement with each other and a strong exterior consistency with measurements acquired from the diagnostic train, thus confirming the reliability of the used methodologies. This investigation further confirms that subsidence in two out of the three test regions in the Konya basin is principally caused by the geological composition of the region, which consists predominantly of clay, and excessive groundwater extraction by the agricultural sector. The soil type significantly influences the railway line’s geometry over time, leading to issues like subsidence and deterioration. The planning of the route should consider this circumstance and make efforts to minimize the presence of uneven and weak ground. Otherwise, the condition would result in ongoing costs due to deterioration and a rise in the frequency of maintenance or repairs for the railway component. However, constant and long-term ground subsidence in any location can cause significant damage to underground facilities, buildings and highways, with the exception of railways. This condition would greatly benefit the municipalities and other relevant authorities and should be taken into consideration.
The cumulative subsidence charts in Figure 11, which were created for study areas #1 and #2, demonstrate that InSAR data are a reliable method for detecting vertical levelling issues in railway lines. For instance, if the data obtained from the diagnostic train indicate any changes in the line’s movement, such as a sudden decrease or increase, it might potentially affect the entire or a significant portion of that region. However, it is impossible to figure out this impact using only the diagnostic train, as the train operates on a single track and the track width has a standard gauge width of 1435 mm, which is common typical in Türkiye and also in many other countries. Performing maintenance work by removing all the rails and consistently adding ballast reinforcement is ineffective in this scenario, as the ground would persistently shift. The primary source of this information is derived almost entirely from InSAR data. The technology is crucial for monitoring railway lines due to its significant advantages in this context. In a different scenario, similar to the third location, the diagnostic train encounters obstacles and challenges that hinder its ability to effectively conduct measurements within a station. These restrictions include the station being used as a parking lot and the presence of several switches, among other issues. An advanced InSAR remote sensing method plays a crucial role in such places. Since these measurement systems are fundamentally different in their nature and purpose, they cannot replace each other, but they can support one another in order to extend the observation of deformation over a larger area and increase the frequency of observations. Given the range of possibilities it provides and the strong correlation between SAR analyses and diagnostic train measurements, with an average correlation coefficient of 0.8 and a maximum of 0.93, along with an average standard deviation of the datasets of less than 3.4 mm/year, it should be considered as an additional observation technique. Furthermore, the broad observation required before the construction of the line is exceedingly challenging and nearly unattainable using alternative terrestrial measuring methods. This technique also yields substantial cost and time advantages.
The expected impact of our suggested approaches for monitoring and identifying instability in railway infrastructure is to improve the early identification of rail fatigue for rail asset management. This development is anticipated to result in the augmentation of the sustainable safety of rail transport. Based on information evaluated and reported by the Turkish State of Railways Maintenance Department, it has been determined that 73% of the regional railway lines and engineering structures in Türkiye are more than 70 years old. Hence, employing an InSAR remote sensing approach for assessing such old structures will be highly advantageous for the purpose of monitoring their condition and implementing precautionary measures as needed. In conclusion, the Copernicus Programme’s Sentinel-1 missions are highly valuable and practical for obtaining cost-effective data to monitor railways and their related facilities. These missions offer free and open-access data, further emphasizing their significance.

Author Contributions

Conceptualization, G.K. and Z.C.; methodology, G.K. and Z.C.; software, G.K.; validation, G.K. and Z.C.; formal analysis, G.K.; investigation, G.K.; resources, G.K.; data curation, G.K.; writing—original draft preparation, G.K.; writing—review and editing, G.K. and Z.C.; visualization, G.K.; supervision, Z.C.; project administration, Z.C. All authors have read and agreed to the published version of the manuscript.

Funding

This research received no external funding.

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

Sentinel-1 satellite data were obtained from The Copernicus Open Access Hub. CSK satellite data and diagnostic train raw data are private sources and not publicly available data, as they are only used by the relevant departments of the Turkish State of Railways.

Conflicts of Interest

The authors declare no conflicts of interest.

Appendix A

Infrastructures 09 00152 g0a1
Figure A1. Map of Ankara–Konya High-Speed railway showing the study areas.
Figure A1. Map of Ankara–Konya High-Speed railway showing the study areas.
Infrastructures 09 00152 g0a1

References

  1. Luo, Q.; Perissin, D.; Lin, H.; Li, Q.; Duering, R. Railway subsidence monitoring by high-resolution INSAR time series analysis in Tianjin. In Proceedings of the 19th International Conference on Geoinformatics, Shanghai, China, 24–26 June 2011. [Google Scholar]
  2. Bamler, R.; Hartl, P. Synthetic aperture radar interferometry. Inverse Probl. 1998, 14, R1. [Google Scholar] [CrossRef]
  3. Koohmishi, M.; Kaewunruen, S.; Chang, L.; Guo, Y. Advancing Railway Track Health Monitoring: Integrating GPR, InSAR and Machine Learning for Enhanced Asset Management. Autom. Constr. 2024, 162, 105378. [Google Scholar] [CrossRef]
  4. Chang, L.; Dollevoet, R.P.B.J.; Hanssen, R.F. Nationwide Railway Monitoring Using Satellite SAR Interferometry. IEEE J. Sel. Top. Appl. Earth Obs. Remote Sens. 2017, 10, 596–604. [Google Scholar] [CrossRef]
  5. Chang, L.; Sakpal, N.P.; Elberink, S.O.; Wang, H. Railway Infrastructure Classification and Instability Identification Using Sentinel-1 SAR and Laser Scanning Data. Sensors 2020, 20, 7108. [Google Scholar] [CrossRef]
  6. Chang, L.; Dollevoet, R.; Hanssen, R. Railway Infrastructure Monitoring using Satellite Radar Data. Int. J. Railw. Tech. 2014, 3, 79–91. [Google Scholar] [CrossRef]
  7. Wang, C.; Zhang, Z.; Zhang, H.; Wu, Q.; Zhang, B.; Tang, Y. Seasonal deformation features on Qinghai-Tibet railway observed using time-series InSAR technique with high-resolution TerraSAR-X images. Remote Sens. Lett. 2017, 8, 1–10. [Google Scholar] [CrossRef]
  8. Di Martino, G.; Esposito, M.; Festa, B.; Iodice, A.; Mancini, L.; Poreh, D.; Riccio, D.; Ruello, G. Railway Bridge Monitoring with Sar: A Case Study. In Proceedings of the IGARSS 2018-2018 IEEE International Geoscience and Remote Sensing Symposium, Valencia, Spain, 22–27 July 2018. [Google Scholar]
  9. Poreh, D.; Iodice, A.; Riccio, D.; Ruello, G. Railways’ Stability Observation by Satellite Radar Images; IntechOpen: London, UK, 2018. [Google Scholar] [CrossRef]
  10. Zhang, Q.; Li, Y.; Zhang, J.; Luo, Y. InSAR technique applied to the monitoring of the Qinghai–Tibet Railway. Nat. Hazards Earth Syst. Sci. 2019, 19, 2229–2240. [Google Scholar] [CrossRef]
  11. Meng, Z.; Shu, C.; Wu, Q.; Yang, Y.; Fu, Z. Monitoring surface deformation of high-speed railway using time-series InSAR method in northeast China. In Proceedings of the IOP Conference Series: Earth Environmental Science, Changchun, China, 11–14 October 2020. [Google Scholar]
  12. Wang, Y.; Bai, Z.; Zhang, Y.; Qin, Y.; Lin, Y.; Li, Y.; Shen, W. Using TerraSAR X-Band and Sentinel-1 C-Band SAR Interferometry for Deformation along Beijing-Tianjin Intercity Railway Analysis. IEEE J. Sel. Top. Appl. Earth Obs. Remote Sens. 2021, 14, 4832–4841. [Google Scholar] [CrossRef]
  13. Ciampoli, L.B.; Gagliardi, V.; Benedetto, A.; Alani, M.A.; Tosti, F. InSAR analysis of C-band data for transport infrastructure monitoring. In Eleventh International Conference on the Bearing Capacity of Roads, Railways and Airfields; Taylor & Francis: London, UK, 2022; Volume 2. [Google Scholar]
  14. Zhang, J.; Zhu, W.; Cheng, Y.; Li, Z. Landslide Detection in the Linzhi–Ya’an Section along the Sichuan-Tibet Railway Based on InSAR and Hot Spot Analysis Methods. Remote Sens. 2021, 13, 3566. [Google Scholar] [CrossRef]
  15. Polcari, M.; Moro, M.; Romaniello, V.; Stramondo, S. Anthropogenic subsidence along railway and road infrastructures in Northern Italy highlighted by Cosmo-SkyMed satellite data. J. App. Remote Sens. 2019, 13, 24515. [Google Scholar] [CrossRef]
  16. Ding, P.; Jia, C.; Di, S.; Wu, J.; Wei, R. Analysis and Evaluation of Land Subsidence along Linear Engineering Based on InSAR Data. KSCE J. Civ. Eng. 2021, 25, 3477–3491. [Google Scholar] [CrossRef]
  17. Shami, S.; Azar, M.K.; Nilfouroushan, F.; Salimi, M.; Reshadi, M.A.M. Assessments of ground subsidence along the railway in the Kashan plain, Iran, using Sentinel-1 data and NSBAS algorithm. Int. J. Appl. Earth Obs. Geoinf. 2022, 112, 102898. [Google Scholar] [CrossRef]
  18. Chen, B.; Gong, H.; Chen, Y.; Lei, K.; Zhou, C.; Si, Y.; Li, X.; Pan, Y.; Gao, M. Investigating land subsidence and its causes along Beijing high-speed railway using multi-platform InSAR and a maximum entropy model. Int. J. App. Earth Obs. Geo. 2021, 96, 102284. [Google Scholar] [CrossRef]
  19. Shi, M.; Chen, B.; Gong, H.; Li, X.; Chen, W.; Gao, M.; Zhou, C.; Lei, K. Monitoring differential subsidence along the beijing–tianjin intercity railway with multiband SAR data. Int. J. Environ. Res. Public Health 2019, 16, 4453. [Google Scholar] [CrossRef]
  20. Yan, Y.; Yang, Q.; Jia, Z.; Zhang, X.; Dai, H.; Ng, A.H.-M. Integration of Multiband InSAR and Leveling Measurements for Analyzing the Surface Subsidence of Shield Tunneling at Beijing-Zhangzhou High-Speed Railway. Sensors 2021, 2021, 6640077. [Google Scholar] [CrossRef]
  21. Gao, Q.; Crosetto, M.; Monserrat, O.; Palama, R.; Barra, A. Infrastructure Monitoring Using the Interferometric Synthetic Aperture Radar (Insar) Technique. Int. Arch. of Photogramm. Remote Sens. Spatial Inf. Sci. 2022, 43, 271–276. [Google Scholar] [CrossRef]
  22. Luo, Q.; Li, J.; Zhang, Y. Monitoring Subsidence over the Planned Jakarta–Bandung (Indonesia) High-Speed Railway Using Sentinel-1 Multi-Temporal InSAR Data. Remote Sens. 2022, 14, 4138. [Google Scholar] [CrossRef]
  23. Hu, F.; Leijen, F.J.v.; Chang, L.; Wu, J.; Hanssen, R.F. Monitoring Deformation along Railway Systems Combining Multi-Temporal InSAR and LiDAR Data. Remote Sens. 2019, 11, 2298. [Google Scholar] [CrossRef]
  24. Meng, X.; Qi, T.; Zhao, Y.; Dijkstra, T.; Shi, W.; Luo, Y.; Wu, Y.; Su, X.; Zhao, F.; Ma, J. Deformation of the Zhangjiazhuang high-speed railway tunnel: An analysis of causal mechanisms using geomorphological surveys and D-InSAR monitoring. J. Mt. Sci. 2021, 18, 1920–1936. [Google Scholar] [CrossRef]
  25. Wang, H.; Chang, L.; Markine, V. 21 Monitoring of Railway Transition Zones Using Satellite Radar Data. Sensors 2018, 18, 413. [Google Scholar] [CrossRef]
  26. Üstün, A.; Tusat, E.; Yalvac, S. Preliminary results of land subsidence monitoring project in Konya Closed Basin between 2006-2009 by means of GNSS observations. Nat. Hazards Earth Syst. Sci. 2010, 10, 1151–1157. [Google Scholar] [CrossRef]
  27. Comut, F.C.; Üstün, A.; Lazecky, M.; Perissin, D. Capability of detecting rapid subsidence wıth Cosmo Skymed and Sentinel-1 dataset over Konya city. In Proceedings of the 2016 European Space Agency Living Planet Symposium, Prague, Czech Republic, 9–13 May 2016. [Google Scholar]
  28. Üstün, A.; Tusat, E.; Yalvac, S.; Özkan, İ.; Eren, Y.; Özdemir, A.; Şimşek, F.F. Land subsidence in Konya Closed Basin and its spatio-temporal detection by GPS and DInSAR. Env. Earth Sci. 2015, 73, 6691–6703. [Google Scholar] [CrossRef]
  29. Caló, F.; Notti, D.; Galve, J.P.; Abdikan, S.; Görüm, T.; Pepe, A.; Şanli, F.B. DInSAR-Based Detection of Land Subsidence and Correlation with Groundwater Depletion in Konya Plain, Turkey. Remote Sens. 2017, 9, 83. [Google Scholar] [CrossRef]
  30. Sireci, N.; Aslan, G.; Cakir, Z. Long-term spatiotemporal evolution of land subsidence in Konya metropolitan area (Turkey) based on multisensor SAR data. Turk. J. Earth Sci. 2021, 30, 681–697. [Google Scholar] [CrossRef]
  31. Orhan, O. Monitoring of land subsidence due to excessive groundwater extraction using small baseline subset technique in Konya, Turkey. Env. Monitor. Assess. 2021, 193, 1–17. [Google Scholar] [CrossRef] [PubMed]
  32. Ankara-Konya High-Speed Track. Available online: https://www.tcdd.gov.tr/projeler/ankara-konya-yuksek-hizli-tren-projesi (accessed on 2 September 2019).
  33. Orhan, O.; Kırtıloglu, O.S.; Yakar, M. Konya kapalı havzası obruk envanter bilgi sisteminin oluşturulması. Geomatik 2020, 5, 81–90. [Google Scholar] [CrossRef]
  34. Gundogdu, B.; Erdogan, H.; Oktar, O. Investigation of Disasters with Different InSAR Methods. Turk. J. Geosci. 2021, 2, 33–45. [Google Scholar] [CrossRef]
  35. Ozdemir, A. Sinkhole susceptibility mapping using logistic regression in Karapınar (Konya, Turkey). Bull. Eng. Geol. Environ. 2016, 75, 681–707. [Google Scholar] [CrossRef]
  36. Chaabani, A.; Deffontaines, B. Application of the SBAS-DInSAR technique for deformation monitoring in Tunis City and Mornag plain. Geo. Nat. Hazards Risk. 2020, 11, 1346–1377. [Google Scholar] [CrossRef]
  37. International Union of Railways (UIC). UIC Code-719R. Available online: https://www.uic.org/com/IMG/pdf/uic_719_722.pdf (accessed on 17 May 2022).
  38. Roger-800 Measuring Train from Mermec Group. Italy. Available online: https://www.mermecgroup.com/measuring-trains-e-systems/recording-cars/533/roger-800.php (accessed on 15 November 2019).
  39. Crosetto, M.; Monserrat, O.; Cuevas-Gonzalez, M.; Devanthery, N.; Crippa, B. Persistent scatterer interferometry: A review. ISPRS J. Photogramm. Remote Sens. 2015, 115, 78–89. [Google Scholar] [CrossRef]
  40. Ferretti, A.; Prati, C.; Rocca, F. Permanent scatterers in SAR interferometry. IEEE Trans. Geosci. Rem. Sens. 2001, 39, 8–20. [Google Scholar] [CrossRef]
  41. Karimzadeh, S.; Matsuoka, M. Remote Sensing X-Band SAR Data for Land Subsidence and Pavement Monitoring. Sensors 2020, 20, 4751. [Google Scholar] [CrossRef] [PubMed]
  42. Bamler, R. Principles of synthetic aperture radar. Surv. Geophys. 2000, 21, 147–157. [Google Scholar] [CrossRef]
  43. Pepe, A.; Calò, F. A Review of Interferometric Synthetic Aperture RADAR (InSAR) Multi-Track Approaches for the Retrieval of Earth’s Surface Displacements. Appl. Sci. 2017, 7, 1264. [Google Scholar] [CrossRef]
  44. Soergel, U. Radar Remote Sensing of Urban Areas; Springer: Dordrecht, The Netherlands, 2010. [Google Scholar]
  45. Perissin, D. SAR Super-Resolution and Characterization of Urban Targets. Ph.D. Thesis, Politecnico di Milano, Milano, Italy, 2006. [Google Scholar]
  46. Perissin, D.; Rocca, F. High-Accuracy Urban DEM Using Permanent Scatterers. IEEE Trans. Geosci. Remote Sens. 2006, 44, 3338–3347. [Google Scholar] [CrossRef]
  47. Berardino, P.; Fornaro, G.; Lanari, R.; Sansosti, E. A new algorithm for surface deformation monitoring based on small baseline differential sar interferograms. IEEE Trans. Geosci. Remote Sens. 2002, 40, 2375–2383. [Google Scholar] [CrossRef]
  48. Hooper, A. A multi-temporal insar method incorporating both persistent scatterer and small baseline approaches. Geophy. Res. Lett. 2008, 35, L16302. [Google Scholar] [CrossRef]
  49. Ferretti, A.; Fumagalli, A.; Novali, F.; Prati, C.; Rocca, F.; Rucci, A. A new algorithm for processing interferometric data-stacks: Squeesar. IEEE Trans. Geosci. Remote Sens. 2011, 49, 3460–3470. [Google Scholar] [CrossRef]
  50. Lanari, R.; Casu, F.; Manzo, M.; Zeni, G.; Berardino, P.; Manunta, M.; Pepe, A. An overview of the small baseline subset algorithm: A dinsar technique for surface deformation analysis. Deformation and Gravity Change: Indicators of Isostasy, Tectonics, Volcanism, and Climate Change. Pure App. Geophy. 2007, 164, 637–661. [Google Scholar] [CrossRef]
  51. Perissin, D. Interferometric SAR Multitemporal Processing: Techniques and Applications. Remote Sens. Dig. Image Process. 2016, 20, 145–176. [Google Scholar]
  52. Perissin, D.; Wang, Z.; Wang, T. The SARPROZ InSAR tool for urban subsidence/manmade structure stability monitoring in China. In Proceedings of the ISRSE 2010, Sidney, Australia, 10–15 April 2011. [Google Scholar]
  53. Perissin, D.; Wang, T. Repeat-Pass SAR Interferometry with Partially Coherent Targets. IEEE Trans. Geosci. Remote Sens. 2012, 50, 271–280. [Google Scholar] [CrossRef]
  54. SARPROZ Software Manual. 2016. Available online: http://www.sarproz.com (accessed on 10 June 2019).
  55. Perissin, D.; Ferretti, A. Urban-Target Recognition by Means of Repeated Spaceborne SAR Images. IEEE Trans. Geosci. Remote Sens. 2007, 45, 4043–4058. [Google Scholar] [CrossRef]
  56. Lei, L.; Perissin, D.; Qin, Y. Change detection with spaceborne InSAR technique in Hong Kong. In Proceedings of the Geoscience and Remote Sensing Symposium (IGARSS), Melbourne, Australia, 21–26 July 2013; pp. 338–341. [Google Scholar]
  57. Cando Jácome, M.; Martinez-Graña, A.M.; Valdés, V. Detection of Terrain Deformations Using InSAR Techniques in Relation to Results on Terrain Subsidence (Ciudad de Zaruma, Ecuador). Remote Sens. 2020, 12, 1598. [Google Scholar] [CrossRef]
  58. Tamayo Duque, J.S.; Ruiz-Armenteros, A.M.; Ávila Álvarez, G.E.; Matiz, G.; Sousa, J.J. Study of Recent Deformations in the Bogotá Savanna and the City of Bogotá (Colombia) Using Multi-Temporal Satellite Radar Interferometry. Remote Sens. 2023, 15, 5249. [Google Scholar] [CrossRef]
  59. Fárová, K.; Jelének, J.; Kopačková-Strnadová, V.; Kycl, P. Comparing DInSAR and PSI Techniques Employed to Sentinel-1 Data to Monitor Highway Stability: A Case Study of a Massive Dobkovičky Landslide, Czech Republic. Remote Sens. 2019, 11, 2670. [Google Scholar] [CrossRef]
  60. Gezgin, C. The influence of groundwater levels on land subsidence in Karaman (Turkey) using the PS-InSAR technique. Adv. Space Res. 2022, 70, 3568–3581. [Google Scholar] [CrossRef]
  61. Weiss, J.R.; Walters, R.J.; Morishita, Y.; Wright, T.J.; Lazecky, M.; Wang, H.; Hussain, E.; Hooper, A.J.; Elliott, J.R.; Rollins, C. High-Resolution Surface Velocities and Strain for Anatolia from Sentinel-1 InSAR and GNSS Data. Geophy. Res. Lett. 2020, 47, e2020GL087376. [Google Scholar] [CrossRef]
  62. Hu, Z.; Mallorquí, J.J. An Accurate Method to Correct Atmospheric Phase Delay for InSAR with the ERA5 Global Atmospheric Model. Remote Sens. 2019, 11, 1969. [Google Scholar] [CrossRef]
  63. Sajjad, M.M.; Wang, J.; Afzal, Z.; Hussain, S.; Siddique, A.; Khan, R.; Ali, M.; Iqbal, J. Assessing the Impacts of Groundwater Depletion and Aquifer Degradation on Land Subsidence in Lahore, Pakistan: A PS-InSAR Approach for Sustainable Urban Development. Remote Sens. 2023, 15, 5418. [Google Scholar] [CrossRef]
  64. Ruiz-Armenteros, A.M.; Bakon, M.; Lazecky, M.; Delgado, J.M.; Sousa, J.J.; Perissin, D.; Caro-Cuenca, M. Multi-Temporal InSAR Processing Comparison in Presence of High Topography. Procedia Comp. Sci. 2016, 100, 1181–1190. [Google Scholar] [CrossRef]
  65. Ketelaar, V.B.H. Satellite Radar Interferometry; Springer: Dordrecht, The Netherlands, 2009. [Google Scholar]
  66. EN13848-5:2008; Railway Applications—Track—Track Geometry Quality Part 5: Geometric Quality Levels. European Committee for Standardization: Brussels, Belgium, 2008.
  67. D’Aranno, P.J.V.; Di Benedetto, A.; Fiani, M.; Marsella, M.; Moriero, I.; Palenzuela Baena, J.A. An Application of Persistent Scatterer Interferometry (PSI) Technique for Infrastructure Monitoring. Remote Sens. 2021, 13, 1052. [Google Scholar] [CrossRef]
  68. Lazecky, M.; Hlavacova, I.; Bakon, M.; Sousa, J.J.; Perissin, D.; Patricio, G. Bridge Displacements Monitoring Using Space-Borne Xband SAR Interferometry. IEEE J. Sel. Top. Appl. Earth Obs. Remote Sens. 2017, 10, 205–210. [Google Scholar] [CrossRef]
  69. Gagliardi, V.; Tosti, F.; Bianchini Ciampoli, L.; Battagliere, M.L.; D’Amato, L.; Alani, A.M.; Benedetto, A. Satellite Remote Sensing and Non-Destructive Testing Methods for Transport Infrastructure Monitoring: Advances, Challenges and Perspectives. Remote Sens. 2023, 15, 418. [Google Scholar] [CrossRef]
Infrastructures 09 00152 g001
Figure 1. Photograph depicting a section of the Ankara–Konya High-Speed Railways provided by the Gokhan Kizilirmak.
Figure 1. Photograph depicting a section of the Ankara–Konya High-Speed Railways provided by the Gokhan Kizilirmak.
Infrastructures 09 00152 g001
Infrastructures 09 00152 g002
Figure 2. Geological maps: (a) shows the 1st study area; (b) shows the 2nd study area.
Figure 2. Geological maps: (a) shows the 1st study area; (b) shows the 2nd study area.
Infrastructures 09 00152 g002
Infrastructures 09 00152 g003
Figure 3. Roger-800 performing measurements on the Ankara–Konya high-speed railway. The image was provided by Gokhan Kizilirmak.
Figure 3. Roger-800 performing measurements on the Ankara–Konya high-speed railway. The image was provided by Gokhan Kizilirmak.
Infrastructures 09 00152 g003
Infrastructures 09 00152 g004
Figure 4. Photo showing the position of the laser measurement sensors. The image is provided by the Gokhan Kizilirmak.
Figure 4. Photo showing the position of the laser measurement sensors. The image is provided by the Gokhan Kizilirmak.
Infrastructures 09 00152 g004
Infrastructures 09 00152 g005
Figure 5. Illustration showing the working principle of levelling on a diagnostic train.
Figure 5. Illustration showing the working principle of levelling on a diagnostic train.
Infrastructures 09 00152 g005
Infrastructures 09 00152 g006
Figure 6. Displacement diagram of the railway in the line of sight and at multiple passes.
Figure 6. Displacement diagram of the railway in the line of sight and at multiple passes.
Infrastructures 09 00152 g006
Infrastructures 09 00152 g007
Figure 7. Simplified workflow of PS-InSAR processing in SARPROZ© (adapted from [58,59]).
Figure 7. Simplified workflow of PS-InSAR processing in SARPROZ© (adapted from [58,59]).
Infrastructures 09 00152 g007
Infrastructures 09 00152 g008
Figure 8. Image graphs for each time-series data stack: (a) CSK ascending; (b) S1–B/T65 descending; and (c) S1–B /T160 ascending. They show the 2D spatiotemporal baseline (yyyymmdd) spaces. Each point displays a scene, and each line displays an interferogram concerning a single master, which is represented with a red color dot.
Figure 8. Image graphs for each time-series data stack: (a) CSK ascending; (b) S1–B/T65 descending; and (c) S1–B /T160 ascending. They show the 2D spatiotemporal baseline (yyyymmdd) spaces. Each point displays a scene, and each line displays an interferogram concerning a single master, which is represented with a red color dot.
Infrastructures 09 00152 g008
Infrastructures 09 00152 g009
Figure 9. Reflectivity map showing the reference point location, city center and railway with blue colored text from all radar images.
Figure 9. Reflectivity map showing the reference point location, city center and railway with blue colored text from all radar images.
Infrastructures 09 00152 g009
Infrastructures 09 00152 g010
Figure 10. PSC maps and scatter plots: (a) CSK; (b) S1–B/T65; (c) S1–B/T160. PSC maps (red line means the railway) and mean velocity maps for CSK and S-1 analyses in LOS direction (dark blue line represents the 30 km-long railway).
Figure 10. PSC maps and scatter plots: (a) CSK; (b) S1–B/T65; (c) S1–B/T160. PSC maps (red line means the railway) and mean velocity maps for CSK and S-1 analyses in LOS direction (dark blue line represents the 30 km-long railway).
Infrastructures 09 00152 g010
Infrastructures 09 00152 g011
Figure 11. Vertical accumulated subsidence profiles of the railway along the 1st and 2nd study areas: (a) CSK; (b) S1–B/T65; (c) S1–B/T160; and (d) diagnostic train measurement time-series graphs.
Figure 11. Vertical accumulated subsidence profiles of the railway along the 1st and 2nd study areas: (a) CSK; (b) S1–B/T65; (c) S1–B/T160; and (d) diagnostic train measurement time-series graphs.
Infrastructures 09 00152 g011
Infrastructures 09 00152 g012
Figure 12. Accumulated subsidence graphs for the clustered PSs, where blue color represents PSs from CSK, orange color signs PSs from S1–B /T160, and lastly, grey color denotes PSs from S1–B /T65: (a) Location#1; (b) Location#2; (c) Location#3; and (d) Location#4.
Figure 12. Accumulated subsidence graphs for the clustered PSs, where blue color represents PSs from CSK, orange color signs PSs from S1–B /T160, and lastly, grey color denotes PSs from S1–B /T65: (a) Location#1; (b) Location#2; (c) Location#3; and (d) Location#4.
Infrastructures 09 00152 g012
Infrastructures 09 00152 g013
Figure 13. Specialized workflow model.
Figure 13. Specialized workflow model.
Infrastructures 09 00152 g013
Table 1. UIC-719R soil quality classes.
Table 1. UIC-719R soil quality classes.
Quality ClassCharacterization
QS0Unsuitable Material (such as organic soil, quick clay, rock salt and gypsum)
QS1Poor Material (such as chalk, calcareous clay and clay slate, with MD > 40 and LA > 40)
QS2Average Material (like hard rock with 25 < MD ≤ 40 and 30 < LA ≤ 40)
QS3Good Material (like hard rock MD ≤ 25 and LA ≤ 30)
Table 2. Properties of Roger-800 diagnostic train.
Table 2. Properties of Roger-800 diagnostic train.
ParameterValue
Inspection focusTrack and Catenary
Length/Width/Height22.5 m/2.85 m/4.1 m
Weightabout 70 t
Maximum drive speed120 km/h
Track gaugeTypically, 1435 mm or narrow
Measuring systems available:Track measurement and inspection
Catenary inspection and measurement
Clearance gauge measurement
Tunnel wall inspection
Signaling and TLC monitoring
Table 3. SAR datasets used in this study.
Table 3. SAR datasets used in this study.
DatasetNumber of ImagesPolarization ModeAverage Acq. Sampling (Days)Resolution RNG (m) × AZ (m)
CSK20HH383 × 3
S1B/T6588VV125 × 20
S1B/T16054VV145 × 20
Table 4. Attributes of the PSs located on the railway lines in research areas 1, 2 and 3.
Table 4. Attributes of the PSs located on the railway lines in research areas 1, 2 and 3.
# of PSs on the
Railway Line		Mean
Coherence		Mean Std.
Dev. (mm/yr.)
Study Area-1
CSK830.672.32
S1B/T65570.793.04
S1B/T160470.842.62
Study Area-2
CSK1270.920.98
S1B/T65820.872.37
S1B/T160860.901.92
Study Area-3
CSK40.781.75
S1B/T6540.813.08
S1B/T16040.763.34
Table 5. Comparison of correlation coefficients of diagnostic train and InSAR measurements for 1st and 2nd areas.
Table 5. Comparison of correlation coefficients of diagnostic train and InSAR measurements for 1st and 2nd areas.
Study Area-1
S1B/T65S1B/T160Diagnostic Train
CSK0.7480.8700.780
S1B/T65 0.6990.930
S1B/T160 0.690
Study Area-2
S1B/T65S1B/T160Diagnostic Train
CSK0.6340.9140.880
S1B/T65 0.6350.650
S1B/T160 0.870
Table 6. Comparison of correlation coefficients of PSs for the 3rd study area.
Table 6. Comparison of correlation coefficients of PSs for the 3rd study area.
Study Area-3
(a)Joint Location#1(b)Joint Location#2
CSKS1B/T65S1B/T160CSKS1B/T65S1B/T160
CSK 0.9800.990CSK 0.9810.988
S1B/T65 0.984S1B/T65 0.994
(c)Joint Location#3(d)Joint Location#4
CSKS1B/T65S1B/T160CSKS1B/T65S1B/T160
CSK 0.9860.983CSK 0.9800.989
S1B/T65 0.987S1B/T65 0.984
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Kizilirmak, G.; Cakir, Z. Application of PS-InSAR and Diagnostic Train Measurement Techniques for Monitoring Subsidence in High-Speed Railway in Konya, Türkiye. Infrastructures 2024, 9, 152. https://doi.org/10.3390/infrastructures9090152

AMA Style

Kizilirmak G, Cakir Z. Application of PS-InSAR and Diagnostic Train Measurement Techniques for Monitoring Subsidence in High-Speed Railway in Konya, Türkiye. Infrastructures. 2024; 9(9):152. https://doi.org/10.3390/infrastructures9090152

Chicago/Turabian Style

Kizilirmak, Gokhan, and Ziyadin Cakir. 2024. "Application of PS-InSAR and Diagnostic Train Measurement Techniques for Monitoring Subsidence in High-Speed Railway in Konya, Türkiye" Infrastructures 9, no. 9: 152. https://doi.org/10.3390/infrastructures9090152

Article Metrics

Back to TopTop