Technological approaches and applied aspects of 3D mapping of the Trans-Siberian Railway (on the example of the Tarmanchukan tunnel)

Latkin Vadim Aleksandrovich ORCID: 0000-0002-9353-0662 Postgraduate student, Department of Geodesy, Physics and Engineering Structures, Altai State Agrarian University 656049, Russia, Altaiskii krai, g. Barnaul, ul. Molodezhnaya, 29a, kab. 1103/3 latkinvadim@mail.ru Krupochkin Evgenii Petrovich PhD in Geography Associate Professor, Head of the Department of Economic Geography and Cartography, Altai State University  656049, Russia, Altaiskii krai, g. Barnaul, ul. Prospekt Lenina, 61, aud. 502M krupochkin@mail.ru Other publications by this author

Vladimirov Vladimir Nikolayevich Doctor of History Professor, Department of Russian History, Altai State University 656049, Russia, Altai Krai, Barnaul, Lenin Avenue, 61, room 312 vvladimirov@icloud.com Other publications by this author

Introduction

The presented work continues a series of publications on the scientific project "The role of the Trans-Siberian Railway in the development of infrastructure, economy and socio-demographic potential of the eastern regions of Late Imperial Russia" with the support of the Russian Geographical Society. Previously in T. Y. valetova considering the use of open map services (Google, Yandex, Bing, OpenStreetMap, Wikimapia, etc.). Special attention is paid to the work with a map service GoogleEarth. The comparison of available cartographic services with satellite images of the Earth in the context of the tasks being solved is carried out ^[1].

In the article by V.N. Vladimirova and E.P. Krupochkin, it is proposed to create a thematic resource on the history of the Trans-Siberian Railway using modern Web-GIS tools. According to the authors, this allows accumulating the source information collected during the research and presenting it in an accessible and convenient form. Several modes of working with online GIS are proposed - interactive, tabular (with a description of the source database), as well as a mode of editing information in the cloud storage database or in a limited mode of data entry or loading. The authors launched a GIS project on the Transsib with accessible layers generated and edited based on digitization of historical and modern maps of different scales ^[2].

The Tarmanchukan Tunnel is one of seven Khingan tunnels located on the 8140th kilometer of the Trans–Siberian Railway in the Arkharinsky district of the Amur Region between the Tarmanchukan and Kundur-Khabarovsk stations. This is the longest tunnel of the Trans-Siberian Railway, the length of which is 2030 m.

The old single-track tunnel was built in 1916 and mothballed in 2005. The key problem was the difficult hydrogeological conditions that led to significant waterlogging and a tendency to large ice formation. Therefore, in 1990, the construction of a new double-track tunnel began.

The project of the Tarmanchukan tunnel was developed by the branch "Bamtonnelproject" of the design institute "Lenmetrogiprotrans" with the involvement of the subcontractor organization of the institute "Dalgiprotrans". Due to the fact that the construction of the tunnel with a standard period of 5 years stretched for 14 years due to insufficient funding, the technical conditions and approaches to the construction of the tunnel have become outdated. Therefore, the Dalgiprotrans Institute carried out an adjustment of the project. Due to the fact that the geological section was made conditionally when tunneling, a discrepancy between the hydrogeological characteristics of the project was revealed. The water flow into the tunnel was 1.5-2.5 cubic meters/hour. To solve this problem, in 2002, a drainage tunnel was drilled for the entire length of the tunnel with the installation of drainage wells and failures with the main structure.

Based on the actual mining and geological conditions, the Tarmanchukan tunnel was tunneled throughout by the method of the lower ledge with approaches from 1 to 3 meters using the smooth blasting method. Drilling was carried out by a self-propelled drilling rig on the MoAZ-64011 chassis with two manipulators, which allows drilling a face with an area of up to 40 m2 from one position. The permanent lining of the tunnel is made of monolithic reinforced concrete with a horseshoe-shaped internal transverse outline with a reverse arch for the perception of hydrostatic pressure. A tunnel has been constructed in soils with a wide range of properties: so the strength of the rocks ranges from 1 to 12 on the scale of Prof. Protodyakonova ^[3].

Methods and technologies

Turning to the key aspects of the technology we use and the main differences between traditional cartographic models and three-dimensional ones, we note that conventional two-dimensional maps limit human perception to two dimensions. A real revolution in cartography has occurred with the development of such areas as mobile, multimedia and animation cartography. The transition from two–dimensional to three-dimensional maps, which appeared as a result of the formation of a new scientific direction - digital three-dimensional (3D) mapping, deserves special attention.

It is quite difficult to give an accurate definition of three-dimensional maps. They can be interpreted in different ways, as in the case of two-dimensional analogues, since 3D maps can have different shapes and properties, different interfaces and interaction tools, be used in various applications, and also be based on various technical solutions.

It should be noted that currently the potential of 3D mapping is not fully used. The reason for this may be the complexity of high-quality 3D solutions and, as a result, an increase in the time and cost of developing systems. Do not forget about such factors as higher requirements for equipment for 3D systems, as well as the already established traditions and habits of cartographers and older generation map users familiar with traditional maps. The problem of the availability of survey data of sufficient quality for a three-dimensional representation of a particular territory has not been completely solved. Most of the attention is paid to built–up urban spaces, much less to rural areas. Currently, many available products based on 3D maps are far from perfect in interface and display quality ^[4].

The purpose of this work is to build a 3D map of one of the sections of the Trans-Siberian Railway (Tarmanchukan Tunnel) using the Prism3D game engine.

Tasks to be solved:

1. Justification of the technology chosen by the authors, description of the main stages of building a 3D map using spatial information (spatial coordinates).

2. Comparative analysis of the results obtained on the basis of information available on the Internet.

3. Development of instructions for working with the resulting digital product.

Speaking about the relevance of the development and use of three-dimensional maps, it is worth noting that they have a wide range of potential applications in navigation, tourism, urban planning, ecology, landscape design and other spheres of human activity. The presence of three-dimensional maps of the territory where the construction or reconstruction of some objects is planned allows you to visually and in detail perceive a new object, make the necessary decisions quickly enough, as well as speed up the design and approval processes. Visually appealing 3D models are an effective way to communicate planning decisions to the public ^{[4, 5]}.

Compared to two-dimensional maps, three-dimensional maps offer significant improvements and advantages. Everything that surrounds us in reality is a three–dimensional three-dimensional world, so the main advantage of three-dimensional maps is the most reliable representation, a high degree of recognition of the depicted objects of three-dimensional space by the user. This is a consequence of how three-dimensional images are presented to our brain, and since childhood, when we learn to recognize the objects that surround us: houses, trees, cars, and so on.

It is also worth noting the fact that traditional cards are not easy to decrypt. After all, first you need to mentally imagine the displayed area, then translate the symbols depicted using the legend and compare the map with the real world. Accordingly, the process of decryption requires some time and effort, as a result of which many people have difficulties in using two-dimensional maps. 3D maps in this regard are much easier to represent, and they can easily be used for their purposes not only by specialists, but also by ordinary users not related to cartography ^[4].

It is necessary to understand what types of 3D representations exist. In the process of analyzing various literary sources, it was possible to identify two similar concepts: a 3D map and a 3D terrain model, or a digital terrain model (DMM). In our country, these concepts are often synonymous, replace each other, and there is no clear division between them. But attempts have been made in foreign literature to identify differences between them, or rather between disciplines such as cartography and geovisualization. Regarding the types of 3D representations ^{[4, 5]}, we proceed from the fact that, in essence, a 3D map is a product of cartography, and a three-dimensional terrain model is the result of geovisualization. Let's take a closer look at what their features and differences are.

Let's draw an analogy with a geographical globe – a model of our planet. Initially, it is a simple material surface in the shape of a ball, on which either a satellite image of the real surface of the Earth or a 2D map is subsequently "stretched". A three-dimensional model of the terrain is built in approximately the same way, while there are 2 options:

1. Specialists with the help of special equipment scan the surface of the Earth, photograph objects of the terrain from different angles. Then, from the obtained materials with geographical reference data, a ready-made three-dimensional image is obtained in camera conditions, on which everything is accurately depicted, as in reality, only on a reduced scale.

2. In specialized programs (GIS, graphic editors), a cloud of multiple points with a planned and altitude position is built. It turns out a surface, which is then overlaid with a satellite image of the Earth's surface with the geodata of the binding. Thus, first a digital relief model is obtained (DEM, 2.5D representation). Then various three-dimensional objects, modeled in advance, are applied to this surface. They may not exactly repeat their real counterparts, but they may be similar to them. As a result, a digital model of the terrain (CMM, 3D representation) is obtained.

However, regardless of which version of the construction of a three-dimensional terrain model is used, when creating it, there is a need for ready-made photographic materials (surface images, photographs of objects). This indicates that the model is completely (or practically) photorealistic. But photorealism, in addition to the obvious benefits, has a downside – it oversaturates the model with information that is not necessary for analysis and does not allow you to concentrate on more important things. Also, having pictures of the surface and photos of objects in the finished model, it is impossible to make any edits, i.e. work with the image of the surface, or with point objects on it. In addition, the satellite image of the surface itself is flat, there is no volumetric display of its components (for example, the type of herbaceous vegetation, various crops, materials, which can be of great importance in the analysis) ^[5].

Creating a three-dimensional map is much more time-consuming and difficult from the point of view of human strength and capabilities. A three-dimensional map (3D map) is a kind of constructor, where all construction takes place "from scratch". Ready-made terrain images and photos of objects are used only to find out the necessary information, but they are not applied to the map itself. In order to display the terrain surface on a three-dimensional map (along with objects), several important work steps must be performed.

The first stage is preliminary modeling in a special program. In the case of a surface, a template of a small area can be modeled. At the same time, the Earth's surface on the map can be modeled according to altitude data (point cloud, isolines), or even constructed manually in the form of independent slope formation. Individual point objects are also modeled for further placement: trees, buildings, various signs, etc.

The second stage is spatial coordination or geo–linking, which assumes placement on the surface according to mathematical laws in accordance with the real spatial position (i.e. taking into account coordinates, distances and angles).

For an effective representation of reality, it is not necessary to display the situation with photographic accuracy, as is done in 3D models of the terrain. The main thing is to adhere to certain cartographic rules: the mathematical law of transferring (projecting) images and objects in space, symbolization, abstraction, generalization.

Symbolization is the direct use of 3D symbols on the map.  To recognize such symbols, if they are properly designed, no special training or symbols are required, since they reflect real-world objects, information about which is accumulated by the user from previous life experience.

Using the abstraction rule, only objects or properties that have a certain value and are necessary for display are taken into account on the map. Other, non-essential characteristics or items can be neglected. The map may seem somewhat "empty", but in fact it is more effective than a map/model overloaded with information.

Generalization or generalization is the union of certain properties of homogeneous objects. These include water bodies, vegetation, communication routes, real estate, etc. When modeling and mapping, there is no need to take into account the features of each individual object, it is enough that it reflects the main distinctive and important characteristics of its group. For example, when mapping an area to display grassy vegetation, it is not necessary to reflect individual types of grasses of the same type, their diversity, but can be combined into a common – steppe vegetation ^[5].

An important aspect of the issue in three-dimensional mapping is the use of specialized software. Already known programs with 3D representation capabilities include GIS Panorama, MapInfo Professional (Virtual Mapper), ArcGIS (Arc Scene), ArchiCAD, etc. ^{[6, 7, 8]}. During the analysis, a subjective opinion was formed about the significant disadvantages of the mentioned products, among which it should be noted the poor quality of models and displays, the lack of a wide variety of types and shapes of objects in certain groups (vegetation, road communication, buildings, etc.) or the absence of some types at all ^[4].

In this regard, there was a need to qualitatively improve the final result of three-dimensional maps. An idea arose to use game engines for this purpose. It cannot be called radically new ^{[9, 10]}, however, in Russia, as a rule, the main attention is paid to specialized GIS, some examples of which were given above. Using the Internet capabilities, a search was carried out for a suitable mapping tool. The Prism3D game engine was selected and analyzed, which is designed to work with large open spaces. In 2012, a computer game was developed based on it, where the creation of the game world was carried out in a graphic editor. This engine was taken into account in order to create a three-dimensional map ^[4].

Among the main areas of the editor's interface, the following can be distinguished (marked with numbers):

1. Main menu – provides access to all the features of the editor.

2. The toolbar is a set of tools for creating a map and certain control mechanisms.

3. Status bar – displays the current position of the mouse cursor when it is moved in the map window (in the editor's own three-dimensional coordinates), information about the map object when it is selected, distances between certain points.

4. Map area – directly mapping the terrain.

All the main work on creating a 3D map was to place three-dimensional models of the terrain surface and objects. Before starting mapping, it was necessary to do some preparatory work – to determine the scale of the map and adjust the coordinate system of the editor.

Before creating any map, you need to determine its scale. It should be noted right away that on a three-dimensional map, the concept of scale is somewhat different from the generally accepted understanding. Traditionally, the scale of a 2D map is interpreted as "the degree of reduction of objects on the map relative to their size on the earth's surface" ^[11]. That is, for example, a scale of 1:10000 means that the size of an object on the map is 10,000 times smaller than its actual size on the ground.

Such an understanding of scale is not suitable for a three-dimensional map because of the possibility of free movement in space (approaching objects or moving away from them), as well as due to the application of the rules of abstraction and generalization. Therefore, we define the scale of the 3D map being created as the degree of correspondence of the distances measured on the map to the actual (real) distances on the ground.

The coordinate system was set up as follows. The location of any object on the map is determined by spatial data – distances, angles, coordinates. In three–dimensional cartography, a third is added to the two traditional X and Y coordinates - Z or H (height). There are many different coordinate systems, and before you start placing objects on the map, you need to familiarize yourself with the coordinate system of the graphic editor.

Since the editor was not originally intended for scientific purposes, the coordinate system in it is specific. Figure 1-2 shows the order of coordinates in the editor window, as well as the direction of the X and Y axes (top view). 

Fig.1. The order of coordinates in the editor

Fig.2. X and Y direction

This coordinate system is rectangular. However, it is organized somewhat differently: as if it is mirrored along the X axis, which is directed downward. In order to plot objects on the created map using real coordinates, using the Internet resource "Google Maps", which uses the WGS-84 coordinate system, it is necessary to calculate formulas for bringing the coordinates of the editor to real coordinates.

The starting point of the map (0.00; 0.00) is the north-western entrance to the Tarmanchukan tunnel (49.171850°; 130.680730°), located "on the 8140th kilometer of the Trans-Siberian railway in the Arkharinsky district of the Amur region ..." ^[12]. To switch from geographical coordinates to rectangular coordinates, you need to know what 1° is in meters. Knowing the geographical coordinates of two points first on the X axis, then on the Y axis, as well as the distances between the points, you can get a value of 1 ° in meters along the meridian, as well as along the parallel.

To obtain this value, the following actions were performed (similar for meridian and parallel):

1. Two points are taken, their geographical coordinates are determined using Google Maps GIS, the distance between them is measured using Google Earth GIS.

2. Taking the difference in the coordinates of two points in degrees and multiplying by the desired value of 1 ° in meters, you can get the distance between the two points. Accordingly, already knowing the distance, you can calculate the length of 1 ° along the meridian and parallel.

As a result , the following values were obtained:

Length of 1 ° along the meridian = 113450 m.

Length of 1° parallel = 72775 m.

Of course, if we take into account the shape of the Earth, the values will differ at different latitudes. But when mapping a local object occupying a relatively small area, the latitude difference will be small, so the changes can be ignored and the obtained values of length 1° in meters along the meridian and parallel can be taken as constant.

Taking into account the geographical coordinates of the starting point, the values of the length of 1 ° along the meridian and parallel in meters, as well as the directions of the axes in the graphical editor (Fig. 2), formulas for the transition from one coordinate to another were proposed.

– Transition from real coordinates to map coordinates:

X maps, m = (49,171850° – X real. °) * 113 450 m (1)    

Y of the card, m =(Y real. ° – 130.680730°) * 72,775 m (2)

– Transition from map coordinates to real ones:

X real., ° = 49.171850 ° –

Y real., ° = 130.680730 ° +

To create a three-dimensional map , the following source materials were used:

1. Data from space and ground surveys of the terrain on the Internet.

2. Technical means (personal computer with component elements), necessary software (Prism3D graphic editor).

3. Electronic files of 3D game maps for obtaining the necessary three-dimensional models of objects (with the indication of the authors) ^[13-16].

It is important to note that in order to display the actual state of three-dimensional objects, verification using their own sources was not carried out. Ready-made models of terrain and objects located on the Internet resources listed above were used for mapping. At the same time, the author's game cards themselves were not used for personal purposes, only models of individual objects. 

Results

The three-dimensional models of point objects, the constituent elements of linear and areal objects and surfaces were displayed using the above formulas (1) and (2). Let us consider this process sequentially.

1. The placement of the main highway – the railway (Fig. 3). Is carried out taking into account the number and structure of the tracks. The width of the entire structure, the type of coating, and the height at certain points of the railway track are adjusted.

Fig. 3. Placement of a railway with an earthen bed

2. Placement of objects within the width of the road (Fig. 4).

The necessary objects are placed within the width of the railway track with the roadbed: tunnel, supports and pillars, signs, moving objects, buildings and structures, etc. The height of objects is indicated either by numerical data in the properties of objects, or manually adjusted by moving along the Z axis (vertically upwards). If necessary, the object can be expanded along three axes, as well as dimensions can be changed in three dimensions: length – height – width in certain proportions, for example, 1:1:1, 1:2:3 etc .

Fig. 4. Placement of objects within the width of the railway track and the roadbed

3. Placement of secondary roads (Fig. 5) and water objects.

At this stage, it is necessary to display other communication routes on the map, for example, various highways with a part of the roadbed with a certain type of vegetation. The height of certain road points is set similarly to clause 2.1. If available and necessary, the surfaces of water bodies are also placed at this stage. The type of surface, vegetation on it, the height at each point of the object (if it is a river/stream) is configured.

Fig. 5. Placement of secondary roads

4. Placement of terrain surfaces (Fig. 6).

There is a creation of areas outside the main highway, between secondary roads, from different sides of water surfaces (if available). The type of coverage of the created sections is set, a certain relief is formed at certain points, or using ready-made forms proposed by the editor.

Fig. 6. Placement of terrain surfaces

5. Placement of various objects on the ground (Fig. 7).

Represents the final stage of work, individual objects (power lines, transformers, buildings, fences, etc.) are plotted according to coordinates. Mass objects (forest, numerous shrubs) are displayed as a group at once in the properties of surfaces (roadbed), and at the same time the types of groups of objects, different distances, density of objects, their sizes, etc. are taken into account.

Fig. 7. Placement of objects on the ground

Thus, as a result of the work done, a three-dimensional map of the territory of the Arkharinsky district of the Amur region with the Tarmanchukan tunnel was created. The overall dimensions of the map are in length (from northwest to southeast) 2500-2700 m, in width (from southwest to northeast) 1300-1800 m. Figure 8 shows several photos of the tunnel.

8. Photos of the Tarmanchukan tunnel: 1 – before entering the tunnel (southeast) ^[17], 2 – entering the tunnel (southeast) ^[18], 3 – inside the tunnel ^[19], 4 – leaving the tunnel (northwest) ^[20]

The materials are taken from Internet resources, links to which are given in the explanatory inscription to the figure. Photographs are the most important data for mapping, because they provide information about the appearance of surfaces and objects visible on satellite images, allow them to be recognized and most accurately recreated on a three-dimensional map.

Fig. 9-14 shows the results of mapping the terrain surface and placed objects in the environment of the Prism3D editor ^[21]. Fig. 9-11 shows the south-eastern entrance to the tunnel from different angles, Fig. 12-14 refer to the north-western exit.

Fig. 9. South-eastern entrance to the tunnel (1)

Fig. 10. South-eastern entrance to the tunnel (2)

Fig. 11. South-eastern entrance to the tunnel (3)

Fig. 12. Northwest entrance to the tunnel (1)

Fig. 13. Northwest entrance to the tunnel (2)

Fig. 14. Northwest entrance to the tunnel (3)

The research results reflect the peculiarities of the construction and operation of the highway, demonstrating the complexity of the natural-geographical and landscape conditions of laying individual sections of the Trans-Siberian Railway. Technologically, this will be presented on a website with a built-in Web-GIS system and a 3D interface (it is under development on the website of the Faculty of History of Moscow State University). 

Conclusion

In the course of research and practical testing of the proposed technology, a 3D map of one of the sections of the Trans-Siberian Railway (Tarmanchukan tunnel) was created. A comparative analysis of existing 3D terrain representation technologies used for scientific purposes has shown that they have weaknesses, such as poor visualization quality, flat surface mapping, and a meager library of three-dimensional object models. From this point of view, existing and actively developing products in the gaming industry have, in our opinion, great potential in geoinformatics, cartography and historical informatics and should be introduced into scientific and practical activities based on their combination with the mechanisms of scientific analysis of specialized GIS to obtain the highest quality software at the present stage. The proposed three-dimensional mapping technology has a number of advantages over other methods of 3D representation, among which one can distinguish:

– three-dimensional image of a surface with a different type of coating depending on specific natural conditions;

– abstraction from non-essential characteristics of the surface and objects, focus only on the necessary data, which saves the user from unnecessary information, variegation of the image, allows you to concentrate on important details;

– generalization of features and properties, which allows you to uniquely identify the type or type of surface/object using a 3D map, i.e. to make an accurate classification and delineation of distribution areas;

– mapping "from scratch", like a constructor, unlike photographing and scanning the terrain, allows you to pay attention to each particular element that is part of the overall picture, "skip through" all the features of the territory under study.

The possibilities of practical use of the developed model are, in our opinion, as follows:

– perception of vertical information (shapes and sizes of surfaces and objects);

– visualization with the possibility of taking measurements inside the model itself;

– overflight of sections of the simulated territory from different angles; fixing a certain type (viewpoint);

– editing, correction of objects and surfaces on the map when changes have occurred at any time (for example, the inclusion of individual details identified as a result of historical analysis, or their exclusion or clarification);

– creating new sections in a given space and then joining (adding) these sections to the general model;

– remote work – the implementation of the processes of analysis, planning, design remotely, without direct travel to the area.

According to the authors, an interesting continuation of this work will be the reconstruction of the old Tarmanchukan tunnel with the construction of three-dimensional maps and a 3D model of the tunnel in the form in which it was from the moment of construction (1916) to its conservation. A comparative analysis of three-dimensional representations with their placement on the site will be useful for historians, local historians, travelers, etc. categories.

As a result of the presented work, the most realistic map of one of the most famous tunnels in the world – Tarmanchukansky - was obtained. At the same time, the geographical features of the relief and landscape are fully reflected on the 3D map itself, access and viewing of which will soon be possible on the website of the Faculty of History of Lomonosov Moscow State University.