DITM: “tf1732_c014” — 2004/10/25 — 12:49 — page 285 — #1 CHAPTER 14 Applications of Digital Terrain Models DTMs have found wide applications since their origin in the late 1950s, in various disciplines such as mapping, remote sensing, civil engineering, mining engineering, geology, geomorphology, military engineering, land planning, and communications (Catlow 1986; Petrie and Kennie 1990; Li and Zhu 2000; Maune et al. 2001). In this chapter, brief descriptions of various applications will be given. 14.1 APPLICATIONS IN CIVIL ENGINEERING The first application of DTM is in civil engineering, more precisely, highway engin- eering. In 1957, Roberts (1957) proposed the use of DTMs for highway design. One year later, Miller and Laflamme (1958) used the data to set up a cross-section (pro- file) model and coined the concept of DTMs for the first time. Thereafter, Roberts and his colleagues at MIT developed the first terrain modeling system. This system could not only interpolate in the sections (profiles), but also calculate the cut-and- fill between sections and provide useful data for engineering design. By 1966, they had been able to provide programs for road design using DTMs, most of which were based on the cut-and-fill calculation. Many techniques originally developed for road design have been applied to construction engineering, such as the design of reservoirs and dams. DTMs have also been widely applied to other related engin- eering such as mining. In this section, applications in road engineering and water conservancy will be described briefly. 14.1.1 Highway and Railway Design The development of a transportation network is complicated, aiming to provide a network to satisfy the needs of society. The design process can be split into steps such assite investigation, route planning and design, earthwork calculation, pavement 285 © 2005 by CRC Press DITM: “tf1732_c014” — 2004/10/25 — 12:49 — page 286 — #2 286 DIGITAL TERRAIN MODELING: PRINCIPLES AND METHODOLOGY (a) (b) (c) (d) Figure 14.1 Cases involved in the design of roads and railways: (a) excavation in the form of cuts (cross section); (b) embankments in the form of fills (cross section); (c) digging in the form of tunnel (profile); and (d) construction in the form of bridge (profile). design, bridge and tunnel design, and so on. DTMs help in route planning and design and earthwork calculation. Due to variations in the terrain, it is unlikely that a road or railway can be construc- ted without any earthwork. In most cases, tunnels and bridges need to be constructed, hills and lowlands modified (Figure 14.1). Earthwork can take the form of excava- tion or the construction of embankments, to carry an elevated highway or railway. Normally in a road or railway design, both cuts and fills will be necessary. Designers make every effort to select a route passing through areas with stable geological conditions, with gentle slopes and small curves to minimize earthwork. Traditionally, such work was done on contour maps. Nowadays, DTMs are widely used for drawing plans, profiles (along the designed central line), and cross sections, for computing the volume of earthwork, for generating perspective views, and even for producing 3-D animation. As various routes are possible for a given project, the aim of the design is to obtain an optimal route. The basic requirements for landscape modification with a TIN are the ability to insert and delete points in a triangulation (see Chapter 5) and to assign to them a particular height. Although it is desirable to have additional tools that are specific to the particular application, terrain manipulation can be done with these alone — other techniques are described in Chapter 15. Figure 14.2 shows a triangulated surface where roadside elevation values have been added to the original terrain, and some of the original data points have been modified. Figure 14.3 shows a 3-D view of this model, which may then be used to evaluate the feasibility of the proposed route. Figure 14.4 shows a highway on the TIN-based DTM with shading. In mining, DTMs have also been used to compute earth volume and to simulate mining progression. 14.1.2 Water Conservancy There are different types of water conservancy projects, such as reservoirs and canals. A canal project is similar to a road project, but there are differences. The major difference is that water cannot naturally flow uphill. Therefore, a canal is normally not allowed to have upward slopes. © 2005 by CRC Press DITM: “tf1732_c014” — 2004/10/25 — 12:49 — page 287 — #3 APPLICATIONS OF DIGITAL TERRAIN MODELS 287 Figure 14.2 A triangulated terrain surface with road lines added. Figure 14.3 A 3-D view of the triangulated terrain surface with road lines. Figure 14.4 A highway designed on TIN-based DTM with shading. © 2005 by CRC Press DITM: “tf1732_c014” — 2004/10/25 — 12:49 — page 288 — #4 288 DIGITAL TERRAIN MODELING: PRINCIPLES AND METHODOLOGY In a reservoir project, the water volume needs to be estimated, the location of the dam determined, various critical water and outlet water levels designed, and drainage planned. Reservoir volume and area are two major features. DTMs can be used to replace traditional contour maps to assist in selecting a site for the dam and estimating water volume. The procedure for the computation of water volume is as follows: 1. Let points (x A , y A ) and (x B , y B ) be the twopointson the dam axis, then theequation for this axis is y = kx +b (14.1) where k = (y B −y A )/(x B −x A ) and b = y A +kx A . 2. Compute the intersection points(x i , y i ) of the damwith the contour linesat different levels, which are derived from the DTM of the reservoir area. 3. Compute the area of the irregular polygon formed by each contour (e.g., the kth contour) and the designed dam: A k = 1 2 (x i+1 +x i )(y i+1 −y i ), k = 1, 2, , m (14.2) 4. Compute the volume between the adjacent two areas (e.g., the kthand the (k +1)th): V = 1 2 (A k +A k+1 ) ×H (14.3) where H is the height difference between the two adjacent areas. 5. Compute the volume contained by the reservoir: V = V (14.4) In the end, the curves showing the relationships between water level and reservoir volume and between water level and water area can also be produced with ease. 14.2 APPLICATIONS IN REMOTE SENSING AND MAPPING DTMs have many applications in remote sensing and mapping, such as topographic mapping (contours), thematic mapping, orthoimage generation and image analysis, map revision, and so on. In this section, only the applications in orthoimage genera- tion and remote sensing image analysis are discussed. 14.2.1 Orthoimage Generation To make images useful as backdrops for other thematic information and base maps, it is desirable that the images have characteristics similar to those of maps. This means that the same scaling, orientation, and projection into a geo-referencing system (e.g., a national geodetic system) should be adopted. To accomplish this, a number of requirements must be fulfilled: 1. All image points should be registered in a geo-referencing system such as a national geodetic (or grid) system. © 2005 by CRC Press DITM: “tf1732_c014” — 2004/10/25 — 12:49 — page 289 — #5 APPLICATIONS OF DIGITAL TERRAIN MODELS 289 2. Each point (pixel) of the resulting image should have the same scale if the ground area is small, or else scale variations should follow a map projection. 3. The relative relationships between features should also be retained. Remote sensing images, either satellite or aerial images, do not have such good characteristics due to the distortions caused by the imperfections of camera or scanner systems, the instability of platforms (tilts and flying height variations), atmospheric refraction, the earth’s curvature, and terrain height variations. The two most serious factors are the instability of the platform and terrain height variations. Therefore, geometric rectification is required. To rectify the images, the relationship between the image and ground points needs to be established. For aerial photography, relationships were discussed in Chapter 3 and expressed by Equation (3.3). For scan image, Equation (3.3) can still be used but only for each scan line. Therefore, it is not practical to use Equation (3.3) to model scan images, because there are normally thousands of scan lines in a frame, and this results in too many unknowns. In practice, a polynomial (normally second or third order), as listed in Table 4.1, is used to approximate geometric transformation models. A few control points from both the image and the ground are measured as reference points to solve the coefficients of this model. Then, all points on the image can be transformed to the ground. This is geometric rectification in which distortions are corrected, minimized, or redistributed. However, the distortion caused by terrain variation is still there. To remove this distortion, as shown in Figure 14.5, a DTM is required. In Figure 14.5(a), the distortion caused by relief is shown. The ground point A has a height z over a reference datum. This causes a displacement “aa” on the image. In the case of rectification, if the height of each pixel on the image, for example, “a” in Figure 14.5, is found from the DTM, then a correction can be applied. The determination of the ground height is done by an iterative process (Albertz et al. 1999), as shown in Figure 14.5(b). A′ A O n a′ B N a z H H–z f (a) (b) X 2 X 4 X 5 X 3 X 1 z 1 z 2 z 4 z 5 z 3 Iteration of the ground coordinates O O a Figure 14.5 Image distortion due to relief and its correction: (a) image distortion caused by relief and (b) intersection of ground surface with light ray. © 2005 by CRC Press DITM: “tf1732_c014” — 2004/10/25 — 12:49 — page 290 — #6 290 DIGITAL TERRAIN MODELING: PRINCIPLES AND METHODOLOGY A B A B Sun Figure 14.6 Effect of topographic variation on image brightness. Table 14.1 Removal of Topographic Effect with Band Ratio Image Brightness on Image Brightness on Slope Facing Away from Sun Slope Facing Sun Unit Red Band NIR Band Red/NIR Red Band NIR Band Red/NIR A 20 30 0.67 40 60 0.67 B 30 60 0.50 50 100 0.50 14.2.2 Remote Sensing Image Analysis In Chapter 12, shading was used to produce a vivid representation of terrain surface. It was argued that the surface reflectance is different if the slope and aspect are different. This means that the brightness of image pixels is also affected by the slope and aspect of the terrain surface. Figure 14.6 shows such effects. However, the image from all the bands will suffer from the same effect for the same area. Therefore, in remote sensing, band ration is widely used to remove such effects. In Figure 14.6, there are two types of land cover, A and B. The sunlight is from the lower-right corner. Therefore, the surface on the right side will appear to be brighter on the image. Table 14.1 gives a possible example. The ratios of the red and the near infrared (NIR) bands are the same although the absolute differences are very different. With a DTM, the slope and aspect map of the terrain surface can easily be produced, as discussed in Chapter 13; thus, the topographic effect can then be removed. In the end, a more reliable analysis could be made from the image after the removal of topographic effects. 14.3 APPLICATIONS IN MILITARY ENGINEERING 14.3.1 Flight Simulation Pilot training is difficult, costly, and sometimes dangerous. It is natural to think about simulation so that the pilot can sit down in front of a special device to learn how to control an airplane. In addition to pilot training, flight simulation can also be used for mission planning and rehearsal. © 2005 by CRC Press DITM: “tf1732_c014” — 2004/10/25 — 12:49 — page 291 — #7 APPLICATIONS OF DIGITAL TERRAIN MODELS 291 Figure 14.7 Virtual battlefield environment simulation (You 1991). In simulation, the DTM plays an important role. 3-D rendering techniques are employed to simulate the terrain, often with LOD techniques. Textures and other attributes can also be mapped onto the DTM surface to generate realistic scenery. To simulate scene changes while flying, the “fly-through” technique is used. DTMs can also be used to guide cruise missiles. This is done by matching the DTM surface stored in the computer with the real world sensed by the detectors on board the cruise missile. 14.3.2 Virtual Battlefield The virtual battlefield is a simulation of a potential battlefield generated in computers, which allows people to be involved. Battlefield simulation provides a dynamic and stereo environment, which can be used to recapitulate the battle, evaluate the results, and gain experience. DTM is used to simulate the battle environment. Figure 14.7 is an example of the virtual battlefield environment. A number of parameters can be derived from the DTM for the battlefield simu- lation, such as intervisibility, shields of the landform, exposed distance of a moving unit, the closest shielding distance to the target, and accessibility of the battle field. 14.4 APPLICATIONS IN RESOURCES AND ENVIRONMENT 14.4.1 Wind Field Models for Environmental Study In climatology, environment, and forestry, to predict the spread of forest fires and pollution, it is necessary to be able to predict the wind direction. To do so, enough information about the wind model of the areas of interest needs to be obtained. A large mountain range exerts dynamic and thermodynamic effects on the atmosphere. The dynamic effect means that with rotation and gravity, mountains force the air to flow like waves at various scales, which results in changes in the wave fronts. The thermodynamic effect means that temperature differences between day © 2005 by CRC Press DITM: “tf1732_c014” — 2004/10/25 — 12:49 — page 292 — #8 292 DIGITAL TERRAIN MODELING: PRINCIPLES AND METHODOLOGY and night in different parts of the mountains result in local influence of hill and valley winds. The direction and velocity of winds also vary with altitude, slope, aspect, and terrain roughness, which results in a complex and unstable wind field in mountain areas. To model this, all the essential topographic parameters can be extracted from DTMs, such as 1. the lowest, highest, and average elevations of each grid cell 2. the lowest, highest, and average elevations of certain square areas delineated with specific conditions 3. the average slope of each cell 4. the percentage of cells containing such terrain features as ridges, valleys, and flat lands 5. the standard deviation of elevations and slopes. 14.4.2 Sunlight Model for Climatology The mountain climate is deeply influenced by terrain variations. These effects result in different climates on the two sides of the mountain range. Also, unique local climates may be formed on any part of a mountainous region because of dif- ferent combinations of altitudes, slopes, and aspects, as well as the shading effect of mountain ridges. When analyzing the mountain climate, one has to consider not only the relatively invariable factors like geographic latitude and average altitude of the region, but also the micro topographic parameters such as the local height differences, aspect, and shading areas. DTMs play an important role in sunlight modeling. The incidence direction of sunlight is defined by the functions of date, time, and the latitude and longitude of the area. The sunlight received by each cell is also dependent on the slope, aspect, and altitude of the cell. To produce a precise sunlight model, the coordinates of grid cells are transformed into latitude and longitude, the angle between the incidence direction of the sunlight and the outward normal of the grid cell is then calculated, the shade status is judged by a hidden surface algorithm of the 3-D perspective, and then the instantaneous sun radiation can be accurately calculated. The sun radiation of a grid cell in 1 day can be obtained by summarizing all instantaneous sun radiation. Similarly, the sun radiation on one grid cell during a month, a season, or a year can be calculated. Of course, the sun radiation at one time period on one slope surface can be obtained by accumulating all the radiation values of those grid cells on the surface. 14.4.3 Flood Simulation The flat areas of river basins are often flooded after heavy rain. Therefore, it is necessary to study flood risks. To do so, potential flood levels and velocity are the two major parameters to be considered. The DTM has been used to simulate floods. In such a simulation, with a given rainfall, the amount of water from different catch- ments can be estimated, as described in Chapter 13. After considering the capacity of the river, the amount of water to be accumulated can be computed. Then, the area © 2005 by CRC Press DITM: “tf1732_c014” — 2004/10/25 — 12:49 — page 293 — #9 APPLICATIONS OF DIGITAL TERRAIN MODELS 293 (a) (b) Figure 14.8 Flood areas simulated with satellite images superimposed. The color plate can be viewed at http://www.crcpress.com/e_products/downloads/download.asp?cat_ no=TF1732. to be flooded can also be estimated. Figure 14.8 shows an example of the flood areas simulated using a DTM, with satellite images superimposed. 14.4.4 Agriculture Management Recently a popular term — precision farming — has come into use. It means that that farmers can control the quantity of water, fertilizer, and pesticides placed on different areas of the farm land, based on the attributes of the land, such as soil type and condition, slope, the condition of the crops, and so on. Slope (as well as aspect) information can be derived from a DTM. Slope is a type of spatial information important to soil erosion. In some developing countries, areas with steep slopes are still farmed, resulting in serious soil erosion. This was also the case, for example, in China. However, in the late 1990s, the Chinese government ordered that no lands with a slope over a certain value should be farmed. In this way, the situation of soil erosion has been improved. 14.5 MARINE NAVIGATION The topographic surface is usually observed above sea level, but clearly the sub-sea terrain surfaceis important in various applications. Terrainmodel construction is often more difficult, as the overall surface form is often not available at the time of sampling. Observations of the sea floor are often made along ships’ tracks, giving a highly anisotropic distribution that requires special interpolation techniques to reconstruct plausible surfaces (Gold and Condal 1995). However, the most important marine application of DTMs is undoubtedly in ship navigation. This is most commonly based on the use of electronic or paper mar- ine charts, but is now being considered for 3-D representations (Gold et al. 2004). Figure14.9showsanavigator’s viewoftheEastLammaChannel,HongKong, together withthesuperimposedchartsymbols.Figure14.10showstheuseofdynamic3-Dsafety contours, highlighting the safe channel for a particular ship’s draught. Figure 14.11 shows how the dynamic intersection of the tidal sea surface with the terrain is used for a collision-avoidance system, with the kinetic Voronoi diagram (derived from TIN) as the basis — collisions are only possible between the generators of adjacent cells. © 2005 by CRC Press DITM: “tf1732_c014” — 2004/10/25 — 12:49 — page 294 — #10 294 DIGITAL TERRAIN MODELING: PRINCIPLES AND METHODOLOGY Figure 14.9 A navigator’s view of the East Lamma Channel. Figure 14.10 Channel safety contours based on sub-sea topography. Figure 14.11 The intersection of the terrain and sea surface: collision detection using kinetic Voronoi diagrams. © 2005 by CRC Press [...]...APPLICATIONS OF DIGITAL TERRAIN MODELS 14. 6 295 OTHER APPLICATIONS In planning and landscape design, visual impact analysis (VIA) is applied to the new designs That is, the designs are superimposed onto a DTM to create a virtual landscape, which is visually analyzed DTMs can also be used for communication network planning Problems such as dead angles and blind areas in site selection... transmitting station can be computed Indeed, as the DTM is a fundamental model of the Earth’s surface, it has applications in all Earth-related sciences However, a complete coverage of these applications lies outside the scope of this book © 2005 by CRC Press DITM: “tf1732_c 014 — 2004/10/25 — 12:49 — page 295 — #11 . #6 290 DIGITAL TERRAIN MODELING: PRINCIPLES AND METHODOLOGY A B A B Sun Figure 14. 6 Effect of topographic variation on image brightness. Table 14. 1 Removal of Topographic Effect with Band Ratio Image. “tf1732_c 014 — 2004/10/25 — 12:49 — page 292 — #8 292 DIGITAL TERRAIN MODELING: PRINCIPLES AND METHODOLOGY and night in different parts of the mountains result in local influence of hill and valley winds Press DITM: “tf1732_c 014 — 2004/10/25 — 12:49 — page 294 — #10 294 DIGITAL TERRAIN MODELING: PRINCIPLES AND METHODOLOGY Figure 14. 9 A navigator’s view of the East Lamma Channel. Figure 14. 10 Channel