Journal of Surveying Engineering

Many existing algorithms for light detection and ranging (lidar) data classification are known to perform reliably; however, the automation of the classification of complex urban scenes is still a challenging problem. In this paper, two classification algorithms based on spatial autocorrelation statistics, such as the Local Moran’s I and the Getis-Ord Gi*, are proposed. These autocorrelation statistics are computed over sample urban areas, including complex terrain with diverse building characteristics. The proposed autocorrelation-based algorithms are applied to airborne lidar point clouds over the complex urban areas to generate highly accurate digital elevation models (DEMs) and classify the lidar points as ground and nonground points by using the DEMs. It is also demonstrated that the minimum-based rasterization and slope-based filtering can be integrated to effectively remove outliers from the DEMs. The test results showed that the autocorrelation-based algorithms produce high-level assessment of overall classification accuracy and Cohen’s kappa index as well as a low level of total errors in complex urban scenes.

The moon is the brightest celestial body in the night sky, and it is of some value when determining the astronomical azimuth on a cloudy or foggy night or in a light-polluted city when there is no visible star. Based on the edge observations of the apparent moon using a total station, a global positioning system (GPS) receiver, a portable computer, and some appropriate software and instrumentation, an innovative method for determining the center of the apparent moon, is discussed in this paper, and calculation formulas are deduced in detail. Second, errors in the astronomical azimuth determination are quantitatively analyzed. Finally, two practical tests are performed in the central plains of China, and the results indicate that the difference between azimuth determination by Polaris and the moon is up to 7″ from two field tests.

A multiresolution wavelet analysis of gravity datasets obtained from satellite, airborne, and terrestrial platforms was conducted to estimate their spectral differences and develop a fused free-air gravity model. In this study, fifth-generation Gravity field and steady-state Ocean Circulation Explorer (GOCE) satellite model (DIRR5 and TIMR5), airborne, and terrestrial gravity data were compared to estimate their differences within various spectral bands using wavelet decomposition. The datasets were combined in a two-dimensional approximation while taking into account their strongest spectral content to derive a fused gravity model that shows improved long and medium wavelengths over those of the original terrestrial gravity dataset. The studies were focused on two distinct areas, namely, the relatively smooth gravity field along the coast of the southeastern United States and the northern Gulf of Mexico and the highly variable gravity field of eastern Alaska. Results show that changes brought by satellite and airborne data in areas where the gravity field is smooth (southern United States) were negligible (0.1–0.2 mGal). However, in areas where the gravity field is more variable and the terrestrial data density is lower, the contribution of the airborne gravity and satellite data was increasingly significant on the order of several milligals. Results indicate future possibilities in predicting the necessity and contribution of airborne data based on the presented regional analysis scheme.

The Atatürk Dam in Turkey, constructed on the Firat River, is one of the world’s largest rockfill embankment dams. It was built to harvest and conserve water and to serve crucial agricultural, social-economic, recreational, and other activities in the Harran region. In this study, geodetic techniques were exploited to assess possible deformations of the embankment structure and add insight on critical dam safety. The geodetic approach included differential, trigonometric, and global positioning system (GPS) leveling observation campaigns over a period of 6.5 years from May 2006 through November 2012. The geodetic control network included 32 reference points and approximately 200 object points located on the surface of the embankment structure. The primary object points were stable concrete platforms embedded on the compacted rock surface, whereas the secondary object points were mounted directly to the compacted rock. All object points were equipped with forced centering mechanisms to support either optical targets, reflectors, or GPS receivers, depending on the type of geodetic measurement campaign. Collected data from different geodetic techniques and results were processed and analyzed. An intercomparison of the results suggests that settlement of the crest reached a maximum of more than 30 cm near the middle of the river. The settlement pattern of the object points on the centerline and upstream side of the crest may suggest shear slump on the clay core with an average of roughly 5 mm/month. The downstream side of the embankment crest appears more stable, with small vertical movement amplitudes (< 10 cm). Fluctuation in reservoir water level, due largely to water management decisions, seems to suggest that variations in water load were unrelated to the embankment surface deformation. On the other hand, the fluctuation in the reservoir water level may have induced variations in the pore pressure of the inner clay core, which in turn may have weakened and caused a slumping on the inner clay core along the centerline and upstream sides of the embankment crest. It is difficult to make definitive claims on the small-amplitude volumetric changes of the downstream embankment surface because the amplitudes of estimated point deformations approached the measurement accuracy limits. However, results indicate that displacements of object points at lower elevation near the thicker base of the downstream embankment surface were more stable than points located higher up, where the embankment is less insulated from mass shifts of the inner clay core. Furthermore, this study supports the notion that GPS leveling can be a viable alternative for monitoring the vertical deformation of large structures when traditional geodetic methods prove too costly.

Only the correct stochastic model can be applied to derive the optimal parameter estimation and then realize the precision global navigation satellite system (GNSS) positioning. The key for refining the GNSS stochastic model is to establish the easy-to-use stochastic model that should capture the error characteristics adequately based on the estimated precisions from the real observations. In this paper, the authors study the GNSS elevation-dependent precision modeling and analyze its impact on the statistic testing involved in the adjustment reliability. With the zero-baseline dual-frequency Global Positioning System (GPS) data, the authors first estimate the elevation-dependent precisions and establish the stochastic models by fitting them with three predefined functions, including the unique precision function and the sine and exponential types of elevation-dependent functions. Three established models are then evaluated by their performance in the overall and w-statistic testing. The results indicated that the GNSS observation precisions are indeed elevation dependent, but this dependence differed from the observation types. The inadequate elevation-dependent model will result in the incorrect statistics and lead to larger wrong decisions, for instance, larger false alarms.

A comprehensive quality-control approach is presented when monitoring motion of tectonic plates using Global Navigation Satellite System (GNSS) measurements of Continuously Operating Reference Station (CORS) networks. The approach was used in a case study in which changes of coordinates with time in a dynamic datum of the newly established CORS network in Western Australia were used to determine the rate of horizontal motion of this part of the Australian Plate. Biweekly data spanning 3 years were used. Quality control started in the preprocessing stage by multipath analysis of all sites and screening of data for outliers, cycle slips, and clock jumps. Possible error sources in the observations were addressed in the processing software. The stability of CORS networks was evaluated by checking changes of their coordinates in a static datum. The computed positions were checked for their consistency, precision, and repeatability. The few problematic data were replaced by better neighboring data sets. The final positions that passed the quality-control process were used for assessment of the Australian Plate movement. Results showed that this plate had a rate of motion in a grid Universal Transverse Mercator (UTM) projection of International Terrestrial Reference Frame ITRF2008 that was 6.84 cm/year with a precision of 2 mm/year and an azimuth of 29.6° on average.

Conformal coordinate transformations very often are not accurate enough to be used for the coordinate transformations between local and global geodetic datums because of the inhomogeneity of the spatial data in different geodetic datums. Therefore, distortion modeling is used to improve the accuracy of the coordinate transformations. This paper presents empirical research on distortion modeling methods used to improve the coordinate transformation process from local to global geodetic datum and vice versa. Various models of distortion shifts in different grid resolutions were computed, evaluated, and compared with the official Croatian model of distortion shifts. Tested modeling methods are inverse distance to a power, kriging, minimum curvature, modified Shepard’s method, natural neighbor, nearest neighbor, polynomial regression, radial basis function, triangulation with linear interpolation, moving average, local polynomial, and least-squares collocation. Distortion model performance was evaluated using summary statistics derived from the back-interpolation for the independent control spatial data set. The present paper proposes the most appropriate method(s) of distortion modeling for the Croatian case and gives brief instructions on possible coordinate transformation improvements of the official Croatian coordinate transformation.

The multibeam echosounding technique is used to provide seafloor mapping at a relatively high-quality level and requires an accurate positioning and orientation system. Integration of a global positioning system (GPS) and an inertial navigation system (INS) is widely used for positioning and orientation control and is performed by using an inertial unit integrated with a real-time kinematic (RTK) GPS solution with two GPS receivers to provide the required accuracy. However, the RTK-based GPS/INS system is presently limited to a distance of approximately 10 km and cannot be used for offshore hydrographic surveys. Recently, precise-point positioning (PPP) technology has shown promise toward the development of an autonomous GPS solution for positioning applications in which only one GPS receiver is used. This paper investigates a real-time PPP-based GPS/INS integrated system position and orientation solution using the international GPS service (IGS) real-time service (RTS) for hydrographic surveys. To meet this objective, the real-time IGS-RTS PPP-based GPS solution is integrated with the INS solution. It is shown that the real-time IGS-RTS PPP-based GPS/INS integration system fulfills International Hydrographic Organization (IHO) special-order 1a, 1b, and 2 requirements at the 95% confidence level and can be used for onshore and offshore hydrographic surveys.

This paper concerns the method of testing reference point stability, which is based on the application of R-estimates, namely, the Hodges–Lehmann estimates (HLEs). The application of R-estimates requires the use of initial residuals, which are computed on the basis of the functional model of observations and the initial values of the parameter vector of such model. The authors considered a case of a leveling network; thus, the main objectives of the paper were, first, to test the way in which the assumed initial values of the parameters affect the R-estimates of vertical displacements of network points and, second, to compare possible ways of assuming the initial point heights. This paper presents some numerical tests that show and illuminate such influences and a comparison of basic methods for practical computation of the initial values of the parameters. Last, the authors found that if the adjusted heights are applied as the initial heights, then the estimation results are satisfactory.

This article presents a new method of arc fitting for railway track realignment. The theoretical foundations are presented, along with a detailed algorithm of the iterative computational process. This method is based on solving a set of linearized pseudo-observation equations. The formulas of the functional model of the fitting problem were derived, and a special form of objective function is proposed. An iterative method for optimization of objective functions is described, and an analysis of the functional model is presented. Specific conclusions are presented based on the results of this analysis.