Nach oben pdf Nothegger, C. & Dorninger, P.: 3D Filtering of High-Resolution Terrestrial Laser Scanner Point Clouds for Cultural Heritage Documentation

Nothegger, C. & Dorninger, P.: 3D Filtering of High-Resolution Terrestrial Laser Scanner Point Clouds for Cultural Heritage Documentation

Nothegger, C. & Dorninger, P.: 3D Filtering of High-Resolution Terrestrial Laser Scanner Point Clouds for Cultural Heritage Documentation

to generate a sufficiently good mesh from the raw point cloud. The registration of point clouds can be real- ized by means of tie points (signalized points or points that can be identified within a scene), relatively (i. e., minimizing the distances of the points belonging to the individual point clouds), or as a combination of both. For regis- tration based on tie points, the achievable ac- curacy increases with the number and geo- metric distribution of given tie points. In many cases, extensive use of signalized points is not feasible (too few natural points and the place- ment of signalized points on the object is pro- hibited), possibly leading to insufficient re- sults. For relative registration, the iterative closest point (ICP) algorithm (R usinkiewicz & L evoy 2001) is commonly used. For that, over- lapping scenes and an approximation of the solution are required. A method, combining tie point observations and relative registration is described by (A kcA & G Ruen 2007). The triangulation of point clouds has been actively studied in computer graphics. Numer- ous algorithms exist that can be used to recon- struct surfaces from noisefree point samples (h oppe et al. 1992, A mentA & B eRn 1999, or D ey & G oswAmi 2003). Some of these algo- rithms have been extended such that they also work with noisy data (D ey & G oswAmi 2006, k oLLuRi et al. 2004, k AzhDAn et al. 2006). They work best when the noise level is low, which makes it necessary to reduce measure- ment noise in situations where noise levels are high. Another problem with the latter two al- gorithms is that they optimize globally, which makes them unsuitable for large datasets. Cur- rently efforts are being made to derive local- ized algorithms (s chALL et al. 2007).
Mehr anzeigen

11 Mehr lesen

Molnár, G., Pfeifer, N., Ressl, C., Dorninger, P. & Nothegger, C.: On-the-job Range Calibration of Terrestrial Laser Scanners with Piecewise Linear Functions

Molnár, G., Pfeifer, N., Ressl, C., Dorninger, P. & Nothegger, C.: On-the-job Range Calibration of Terrestrial Laser Scanners with Piecewise Linear Functions

4.2 Preprocessing of Data The instrument already is equipped with a set of three internal range correction functions: a periodic correction with a wavelength of 60 cm, a polygonal correction function based on range and described as a look-up table, and another look-up table based correction corre- lated with the intensity of the reflected signal. Individual investigations of these three cor- rections have shown that the intensity calibra- tion is sufficient, while the other two correc- tions may not able to cope with the occurring systematic errors properly. Indeed, the influ- ence of the periodic error with a wavelength of 60 cm was reduced but not eliminated suffi- ciently and periodic errors with different wavelengths are still detectable. Moreover, the polygonal distance correction eliminates the systematic errors only partially (d orninGer et
Mehr anzeigen

13 Mehr lesen

Integration of Digital Photogrammetry and Terrestrial Laser Scanning for Cultural Heritage Data Recording

Integration of Digital Photogrammetry and Terrestrial Laser Scanning for Cultural Heritage Data Recording

ƒ Seven-Parameter Transformation. SfM method delivers camera orientations and sparse point clouds, initially in an arbitrary model space. Each synthetic image involved in the SfM process stores 2D-to-3D correspondences between each image pixel or feature and the 3D laser data. This allows an implicit determination of the 3D-to-3D correspondences between the sparse point clouds and the laser data. To introduce scale information to the bundle, a seven-parameter transformation is estimated using the latter 3D correspondences and then applied to the SfM output. This results in having absolute oriented images in relation to the laser data. An alternative method that can increase measurement redundancy is by reprojecting the sparse point clouds onto the synthetic image using equation 4.1. Then, the 3D-to-3D correspondences between the sparse point clouds and the laser data can be determined using the 2D-to-3D correspondences between the latter projected sparse point clouds and the laser data stored in the synthetic image. For the reason that some points will be reprojected from object surfaces that are not covered in the generated image, the geometric relationship of the 3D-to-3D correspondences should be evaluated to remove these wrong points. This can be done using RANSAC filtering scheme based on seven-parameter transformation. Furthermore, an outlier removal process can be applied on the reprojection errors in object space, e.g. using the X84 rule. As mentioned in section 4.2, some dense matching algorithms deliver individual point clouds for almost each image, e.g. the software SURE therefore, it is more convenient to perform first a dense image matching step and then project only the corresponding single point cloud onto the generated image. This can filter out incorrect reprojected points on the generated image. Dense Image Matching. After the estimation of the transformation parameters, the orientation parameters for the camera images are known in the laser scanning coordinate system. These parameters can be used to retrieve dense surface reconstruction information from the images by means of dense image matching methods. The resulting geometry is in the coordinate system of the laser scanner and thus scaled. Supplemental improvement of point cloud registration using ICP is possible.
Mehr anzeigen

164 Mehr lesen

Integration of digital photogrammetry and terrestrial laser scanning for cultural heritage data recording

Integration of digital photogrammetry and terrestrial laser scanning for cultural heritage data recording

 Seven-Parameter Transformation. SfM method delivers camera orientations and sparse point clouds, initially in an arbitrary model space. Each synthetic image involved in the SfM process stores 2D-to-3D correspondences between each image pixel or feature and the 3D laser data. This allows an implicit determination of the 3D-to-3D correspondences between the sparse point clouds and the laser data. To introduce scale information to the bundle, a seven-parameter transformation is estimated using the latter 3D correspondences and then applied to the SfM output. This results in having absolute oriented images in relation to the laser data. An alternative method that can increase measurement redundancy is by reprojecting the sparse point clouds onto the synthetic image using equation 4.1. Then, the 3D-to-3D correspondences between the sparse point clouds and the laser data can be determined using the 2D-to-3D correspondences between the latter projected sparse point clouds and the laser data stored in the synthetic image. For the reason that some points will be reprojected from object surfaces that are not covered in the generated image, the geometric relationship of the 3D-to-3D correspondences should be evaluated to remove these wrong points. This can be done using RANSAC filtering scheme based on seven-parameter transformation. Furthermore, an outlier removal process can be applied on the reprojection errors in object space, e.g. using the X84 rule. As mentioned in section 4.2, some dense matching algorithms deliver individual point clouds for almost each image, e.g. the software SURE therefore, it is more convenient to perform first a dense image matching step and then project only the corresponding single point cloud onto the generated image. This can filter out incorrect reprojected points on the generated image. Dense Image Matching. After the estimation of the transformation parameters, the orientation parameters for the camera images are known in the laser scanning coordinate system. These parameters can be used to retrieve dense surface reconstruction information from the images by means of dense image matching methods. The resulting geometry is in the coordinate system of the laser scanner and thus scaled. Supplemental improvement of point cloud registration using ICP is possible.
Mehr anzeigen

162 Mehr lesen

Potential and Limits of Non-local Means InSAR Filtering for TanDEM-X High-resolution DEM Generation

Potential and Limits of Non-local Means InSAR Filtering for TanDEM-X High-resolution DEM Generation

Figure 10: Scenario with constant reflectivity and coherence. From left to right the results produced by a 5 × 5 Boxcar filter, NL-SAR and NL-InSAR. The true values are delineated by the solid lines, the areas in light blue show 95% percentile of the estimates. The improved noise reduction that non-local filters provide is evident. However the result of NL-SAR shows unacceptable smoothing of the edge.

31 Mehr lesen

High resolution coding of point processes and the Boolean model

High resolution coding of point processes and the Boolean model

The asymptotics of the upper bound are the same as the asymptotics of the quantization error of a (d + 1)-dimensional Poisson point process. The reason for this lies in the construction of the codebook: we use a codebook that first codes a d-dimensional Poisson point process and uses more rate to code the radii of the balls. Hence, intuitively we have d dimensions for the point process and one dimension for the radii. But as the lower bound yields in the case d = 1 the right asymptotics, we conjecture that this yields the right asymptotics as well in the case d > 1. Heuristically, this may be understood by considering the overlaps of the Boolean model. If the radii are quite large with high probability, we have many overlaps in the Boolean model (e.g. some balls may be entirely contained in other balls), and we need not code all points of the Poisson point process. If the radii are that small that we do not have any overlaps, the Boolean model is very close to the d-dimensional Poisson point process, and thus the quantization error asymptotics may be equal.
Mehr anzeigen

128 Mehr lesen

Hacking Cultural Heritage : the Hackathon as a Method for Heritage Interpretation

Hacking Cultural Heritage : the Hackathon as a Method for Heritage Interpretation

communities by appropriating technology. In the 1950s, the first hackers and their technological innovations were driven by a set of principles that influenced many cultures, movements, philosophies, and initiatives that were still to come into existence, such as the Open-source Movement and Wikipedia. Besides giving a historical context to hacking, the chapter also examines the origins of Hackathons as events capable of aggregating communities and intensively focusing their abilities to come up with ingenious solutions to technological problems. Not only hacking, but also Hackathons are based on the same set of principles that enable them to occur, such as free information, decentralization, meritocracy, and the belief that the computer could be used as a tool to change old and create new worlds. Currently, the Cultural Heritage Sector has also appropriated from hacking principles and Hackathons as a way to regaining relevance in a fast-paced and increasingly digitized society, because Hackathons are powerful strategies to advance innovation not only, but also in the Cultural Heritage Sector. Hackathons invigorate the digital and participatory strategies of Cultural Heritage Institutions. Furthermore, these events offer numerous advantages not only to institutions, but also to affiliated communities and the institutions’ audiences. The chapter not only discusses about these advantages, but also provides a closer look at the structures of Hackathons and presents concrete examples of these events and their outcomes in the Cultural Heritage Sector.
Mehr anzeigen

295 Mehr lesen

The contribution of the EXCELSIOR Project for cultural heritage

The contribution of the EXCELSIOR Project for cultural heritage

project is currently integrated within the EXCELSIOR project, whose goal is the development of the Eratosthenes Centre of Excellence (ECoE) . The project focuses on conducting basic and applied research and innovation in the areas of the integrated use of remote sensing and space-based techniques for monitoring the environment. The integration of novel EO, space and ground-based integrated technologies, can contribute to a more sustainable and systematic monitoring of the environment, the timely detection of societal risks/threats and the growth of vital economic sectors. The establishment of the Centre of Excellence in 2020 provides the infrastructure and experts necessary to conduct state-of-the-art research and innovation in the areas of the integrated use of remote sensing and space-based techniques for cultural heritage applications within Cyprus. Strategic Partners of the ECoE, such as DLR and NOA, will also provide capacity building for cultural heritage in order to facilitate state-of-the-art research and applications in cultural heritage.
Mehr anzeigen

7 Mehr lesen

Automatic detection and reconstruction of 2D/3D building shapes from spaceborne TomoSAR point clouds

Automatic detection and reconstruction of 2D/3D building shapes from spaceborne TomoSAR point clouds

Abstract—Modern spaceborne synthetic aperture radar (SAR) sensors, such as TerraSAR-X/TanDEM-X and COSMO-SkyMed, can deliver very high resolution (VHR) data beyond the inher- ent spatial scales of buildings. Processing these VHR data with advanced interferometric techniques, such as SAR tomography (TomoSAR), allows for the generation of four-dimensional point clouds, containing not only the 3-D positions of the scatterer location but also the estimates of seasonal/temporal deformation on the scale of centimeters or even millimeters, making them very attractive for generating dynamic city models from space. Motivated by these chances, the authors have earlier proposed ap- proaches that demonstrated first attempts toward reconstruction of building facades from this class of data. The approaches work well when high density of facade points exists, and the full shape of the building could be reconstructed if data are available from multiple views, e.g., from both ascending and descending orbits. However, there are cases when no or only few facade points are available. This usually happens for lower height buildings and renders the detection of facade points/regions very challenging. Moreover, problems related to the visibility of facades mainly facing toward the azimuth direction (i.e., facades orthogonally ori- ented to the flight direction) can also cause difficulties in deriving the complete structure of individual buildings. These problems motivated us to reconstruct full 2-D/3-D shapes of buildings via exploitation of roof points. In this paper, we present a novel and complete data-driven framework for the automatic (parametric) reconstruction of 2-D/3-D building shapes (or footprints) using unstructured TomoSAR point clouds particularly generated from one viewing angle only. The proposed approach is illustrated and validated by examples using TomoSAR point clouds generated using TerraSAR-X high-resolution spotlight data stacks acquired from ascending orbit covering two different test areas, with one containing simple moderate-sized buildings in Las Vegas, USA and the other containing relatively complex building structures in Berlin, Germany.
Mehr anzeigen

19 Mehr lesen

A Discrimination Method Of Measurement Data Of The Riverbed Landform By Using 3D Laser Scanner

A Discrimination Method Of Measurement Data Of The Riverbed Landform By Using 3D Laser Scanner

Fig.1 is the image figure of the measurement distance, therefore it is not for discriminating aspect of the circumstance. However when 3D laser scanner is utilized actually, required data have to be selected to carry out the required data processing after changing the measurement result to the three-dimensional coordinate system. Although there is no problem when the measurement object is a single material, there are many cases that various objects such as grits, vegetation, water, and river structures are intermingled when measurement of river landform is carried out. For example, the landform eliminating the vegetation of the surface has to be grasped to calculate fluctuation quantity of the river bed, in addition, water area has to be grasped accurately to grasp the water route landform.
Mehr anzeigen

9 Mehr lesen

CHALLENGES IN FUSION OF HETEROGENEOUS POINT CLOUDS

CHALLENGES IN FUSION OF HETEROGENEOUS POINT CLOUDS

In case of symmetries or multiple parallel planes, small rotations and translations do not affect the quality of overlap for most of the points. Thus, the optimization converges to a local minimum of a flat valley in the cost function. Such a case occurs for the scene in fig. 3, where the ground and the parallel rooftop levels constitute most of the correspondences. Therefore in-plane translations and rotations do not affect the overall optimization metric enough to produce a gradient to the optimal alignment. We can solve this problem by extracting planar surfaces (Rabbani et al., 2006) and their boundaries to perform ICP only on boundary points. I.e., we will not consider the large amount of uninformative points for the in-plane rotations and translations. Results for this degenerate case are presented in the sec. 3.1.
Mehr anzeigen

8 Mehr lesen

Chan, T.O., Lichchti, D.D., Belton, D., Klingseisen, B. & Helmholz, P.: Survey Accuracy Analysis of a Hand-held Mobile LiDAR Device for Cultural Heritage Documentation

Chan, T.O., Lichchti, D.D., Belton, D., Klingseisen, B. & Helmholz, P.: Survey Accuracy Analysis of a Hand-held Mobile LiDAR Device for Cultural Heritage Documentation

walked around the target in an open field), the reference objects for the ICP estimation were only the target itself and some sparse vegeta- tion in both tests. Under similar scanning re- ference conditions, the scanning range of the tower (approximately 5 m – 8 m) is higher than that of the previous test (approximately 2 cm – 3 m). As a result, the measurement accuracy of the tower is lower. In addition, the scanning orientation was significantly different com- pared to the previous test because the object was 1 m – 5 m above the scanner. So, there are no horizontal reference objects available at the same height. Even though the estimated mo-del parameters for the prism/pyramid dif- fered more, the plane fitting accuracies for the individual parts were consistent with previous test results (Tab. 9). However, the deviations between interior angles are slightly higher compared to previous tests (Tab. 10). This is likely due to the fact that the walls are much shorter in this case. The mean side length of the tower and the water tank are 0.95 m and 5.76 m, respectively. Therefore, errors in the perpendicular distances (Zebedee) for the tow- er and water tank corresponding to mean an- gular errors, 0.328° and 0.267°, are 5.43 mm and 2.68 mm, respectively. Fig. 13 shows that
Mehr anzeigen

13 Mehr lesen

High Resolution 3D Earth Observation Data Analysis for Safeguards Activities

High Resolution 3D Earth Observation Data Analysis for Safeguards Activities

Abstract. This paper provides an overview of the investigations performed at DLR with respect to the application of high resolution SAR and optical data for 3D analysis in the context of Safeguards. The Research Center Jülich and the adjacent open cut mines were used as main test sites, and a comprehensive stack of ascending and descending TerraSAR data was acquired over two years. TerraSAR data acquisition was performed, and various ways to visualize and analyze stacks of radar images were evaluated. Building height estimation was performed using a combination of ascending- descending radar images, as well as height-form-shadow and height-from-layover. A tutorial on building signatures from SAR images highlighted the sensor specific imaging characteristics. These topics were particularly relevant in safeguards activity with a “small-budget” as only a single image – or a couple - were employed. Interferometric coherence map interpretation allows the detection of traffic on dirt roads. Digital surface models (DSM) were generated from TanDEM-X interferometric data and from optical VHR data. Sub-meter Worldview-2 and GeoEye-1 data was processed into highly detailed DSM with a grid spacing of 1 m, showing building structures. 3D change and volume detection was performed with both optical and radar DSMs. The TanDEM-X DSMs proved useful for volume change detection and computation in mining areas, and DSMs generated from optical satellite data show details on the building level. Virtual 3D fly-throughs were found to be a good tool to provide an intuitive understanding of site structure and might be useful for inspector briefing.
Mehr anzeigen

8 Mehr lesen

Aircraft to Ground Unidirectional Laser-Comm. Terminal for High Resolution Sensors

Aircraft to Ground Unidirectional Laser-Comm. Terminal for High Resolution Sensors

The collimation and coupling system as well as the tracking sensors and filters etc. are mounted on the optical bench. The transmission wavelength of the terminal is 1550 nm with an output power of 100 mW. For the first stage of expansion no Fine Pointing Assembly (FPA) is installed. Therefore just the InGaAs acquisition FPA sensor is installed at the moment and used for acquisition and tracking. The available field of view (FoV) on the optical bench without beam truncation is about 52 mrad. The optical system of the tracking sensor is matched to this large FoV. Therefore a standard inertial measurement unit without high precision requirements can be used to get initial values for pointing toward the ground station prior to the acquisition scan.
Mehr anzeigen

7 Mehr lesen

A novel laser diode wavelength stabilisation technique for use in high resolution spectroscopy

A novel laser diode wavelength stabilisation technique for use in high resolution spectroscopy

3 Laser Diode Temperature Co ntrol and Series Resistance The wavelength stability of a laser diode is vital in TDLAS. In TDLAS the target gas absorption line has narrow linewidth and therefore the detection technique requires excellent wavelength stability (10% better than the linewidth of the absorption line). The emission wavelength and output power of the laser diode are dependent on its operating temperature and injection current. Therefore, it is important to characterise the effects of temperature on laser diodes in order to operate them with a stable output power and emission wavelength. In a conventional laser diode temperature controller, a thermistor sensor-based peltier thermo-electric cooler is used to stabilise the temperature of the laser diode. However, due to the laser diode package design, the thermistor is placed at a distance from the gain chip and does not sense the actual temperature of the laser gain chip. Therefore, the laser diode may drift with change in the ambient temperature. As the laser diode wavelength has large temperature gradient, the laser diode emission wavelength will vary with the change in temperature. Therefore, an alternative method is required to measure the actual temperature of the laser gain chip and to improve the stability of the emission wavelength of the laser diode.
Mehr anzeigen

193 Mehr lesen

Muhle, D., Abraham, S., Wiggenhagen, M. & Heipke, C.: Identifying Correspondences in Sparse and Varying 3D Point Clouds using Distinctive Features

Muhle, D., Abraham, S., Wiggenhagen, M. & Heipke, C.: Identifying Correspondences in Sparse and Varying 3D Point Clouds using Distinctive Features

d k : ∀k = 1…n 2 , d k ∈ D are mapped into an in- teger scalar s k : s k ∈ {1, 2, … 2p}, where p is a parameter of the mapping and inluences the discretisation. Finally, the s k are pooled in an ordered result set S containing only unique values. As shown in section 3.4, the usage of integer values is advantageous as it allows an eficient implementation for the comparison of two distinctive descriptions. The s k are com- puted by applying the mapping T : d k → s k to all entries d k in D. For T the quad tree index (F inKeL & B entLey 1974) is used that recur- sively divides the 2D space into discrete grids and gives an integer index for a two dimen- sional point. The only parameter of the quad tree index is p deining the number of quad- rants used for the partitioning of the two di- mensional space. This parameter controls the loss of accuracy caused by the discretization. For a chosen value of p = 16 the x-axis (radial distances d j ) and the y-axis (angle θ j ) will be partitioned into 2 p/2 = 256 bins. The value of nates. Fig. 1(b) shows that for all neighbours
Mehr anzeigen

11 Mehr lesen

Planar filtering in semiconductor laser diodes with high beam quality and power

Planar filtering in semiconductor laser diodes with high beam quality and power

The vertical design of planar semiconductor lasers is directed to the monomodal property along the vertical direction. The structure in the plane containing the active zone can be designed by the thickness of the waveguide, current distribution, implantation or other means in such a way that the mode emitted at the emission facet is a fundamental mode. The 2-dimensional filtering schemes used for this goal in a 2...4 mm x 200 µm area, for example, are mostly stimulated by the optical filtering technique, the volume holography [1] and by the optical resonator theory.
Mehr anzeigen

4 Mehr lesen

Historical Replication Preserves Cultural Heritage

Historical Replication Preserves Cultural Heritage

We therefore propose a more radical alternative: The major attractions of the most visited places are to be replicated in a suitable place easy to reach by tourists and having few negative effects on locals. In the case of Venice, for example, the Doge Palace, the Saint Marcus Church, the Tower on the Piazza San Marco as well as the square itself, and the Rialto Bridge would be exactly replicated and placed somewhere more suitable on the Italian or Balkan coast. These monuments are the major reason why most tourists want to visit Venice, and it is expected that many tourists would accept the offer, especially as the replicated sites are to be installed with the most modern technology, e.g. having Dodges, and other historical inhabitants, walking around by using holograms. At the same time the replicated sites would offer convenient restaurants and shopping opportunities which tourists value (Yüksel, 2007). The visitors are, of course, aware that they are not in “historical” Venice but this feeling is overcompensated by a more intense historical experience. The replicas suggested can also combine various cities and sites. An example would be “Historical North Italian City States” which could combine the major attractions of cities such as Siena, Pisa, Parma or Piacenza but strongly reducing transportation requirements for visitors engaging in cultural city tourism.
Mehr anzeigen

9 Mehr lesen

Guidelines: Interpretation of European Cultural Heritage in Tourism

Guidelines: Interpretation of European Cultural Heritage in Tourism

The regional affiliation is a distinguishing feature of different culinary systems within a country. Regional cultural areas are used and worked on by the local population due to specific natural conditions (e.g. agriculture). In this way, specific regional culinary networks are created (“foodways”), which, among other things, highlight the special features of the respective region (Long, 2004). Time can also be used as a differentiating feature. Foreign food and beverages from both the past and the future can be considered (Long 2004, p. 26). By means of historical sources, for example, unknown culinary specialties and customs can be experienced (e.g. historically transmitted recipes and cookbooks). Further access is provided by museums that supply knowledge on the production, storage, distribution, preparation and reception of food and beverages, among other things. These knowledge resources can be implemented in the context of tourist attractions (e.g. tastings of historical food and drink or “living history sites”; Long, 2004). In addition to this, however, time can also make reference to foreign culinary specialties with regard to holidays / religious or cultural festivals (food: e.g. gingerbread; preparation methods: e.g. dyeing of Easter eggs; consumer behaviour: e.g. family dinners on certain holidays; Long; 2004). Some cultural factors can influence eating habits within a given culture. For example, religion can affect food preferences through religious bans or certain preparation regulations. For instance, at church festivals and events, certain foods and drinks are exploited to impart knowledge about the religion concerned (Long 2004). Ethos can manifest itself in terms of nutrition through value-driven consumption behaviour (e.g. veganism, organic food). Also, tourists can experience these nutrition forms directly (e.g. vegan restaurants). The social class can also be consulted if individual groups establish culinary networks (Long, 2004). An example is the culinary customs of the working class in the south of the USA or “mountain food” (unusual ingredients such as cornmeal and date plums in the cuisine of the population in the Appalachian, USA; Long, 2004).
Mehr anzeigen

221 Mehr lesen

Fusing meter-resolution 4-D InSAR point clouds and optical images for semantic urban infrastructure monitoring

Fusing meter-resolution 4-D InSAR point clouds and optical images for semantic urban infrastructure monitoring

By detecting the peaks in the derivative of the deformation function, the location of the discontinuities can be detected. Constraint was put on the minimum distance between two peaks representing the minimum length of a railway segment. The detected peaks in the deformation's derivative can be seen in Figure 20(b). The positions of the peaks on the railway are shown as the green dots in Figure 21. Each green dot represents the midpoint of its railway cross-section. In the middle subfigures of Figure 21, the close up view of the two joints in the optical image is provided. As the optical image has limited resolution, we also provide a higher resolution one (7cm ground spacing) in Figure 22. It can be clearly observed that the railway joint shown up as dark lines in the optical image.
Mehr anzeigen

43 Mehr lesen

Show all 10000 documents...