Comparative Study on Executing Topographic Plans Using UAVs

Paunescu C^1*, Marinescu CM² and Paunescu V^{2 1}Department of Geomatics, Bucharest University, Vidra 31, sector 6, Bucharest, Romania
²Department of Geomatics, Vidra 31, Sector 6, Bucharest, Romania
^*Correspondence: Paunescu C, Department of Geomatics, Bucharest University, Romania, Tel: +40722209287, Email:

Received: 21-Sep-2020 Published: 11-Sep-2020, DOI: 10.35248/2469-4134.20.9.279

Methodology

Making the first variant of the digital terrain model

The area where the study was conducted is the commune of Schitu, Giurgiu County, Romania. In the area there have been works for the introduction of the systematic cadastre, respectively the measurement of each property. For this, a DroneZone XF8- CT flight was equipped with a u-Blox NEO8M GNSS receiver and a Sony A7R 35 mm 36 Mpixcamera. Also, the property limits were determined with a total station Leica TS06 type and Leica GS08 Plus GNSS receivers. To georeferencing (scaling) the images taken with the UAV system, 6 control points have been determined, pre-signaled on the ground, before he flight [1-3].

When designing the flight, we took into account the following considerations:

Depending on the products to be obtained, the longitudinal and transverse covers between the images are determined, as well as the height and the speed of flight;

Establishing the orientation of the flight strips;

Determining the best days and times for carrying up the photogrammetric flight;

Analysis of the weather; information is collected from weather stations near the area of interest;

Establishing the final flight route that is handed over to the pilot or operator of the photogrammetric camera.

The longitudinal coverage was of 80% and the transverse coverage of 50% to obtain in the end a trueortophoto. The strips were established according to Figure 1.

Figure 1. Flight strips.

The pre-signaling was marked on the ground either with paint or single-use plates, as in Figure 2.

Figure 2. Pre-signaling the control points on the ground.

 

At the time of flight the flight conditions were optimal: clear sky, temperature over 5°C, wind speed below 2 m/s.

Following the data processing with the Agisoft software, a cloud of points was obtained. In order to obtain the Digital Terrain Model (DTM), we have performed the unsupervised (automatic) classification and then the supervised (manual) classification of the cloud of points in order to establish the points that belong to the ground class [4-6].

By triangulating the points in the ground class, we obtained a solid model known as the Digital Terrain Model (DTM) (Figure 3).

Carrying out the second digital terrain model

A new UAV model has launched on the market, SenseFlyeBee X, equipped with a TRIMBLE BD93 type GNSS RTK receiver, a SenseFly GeoBase base and a SenseFlyAeria X 24MPix camera. From the description that the seller made, it turned out that there was no need for control points (landmarks) for georeferencing [7-9].

We decided to do a comparative study, to convince ourselves that the newly developed system has the performances described by the seller [10].

Thus, we redo the flight we had made with the first type of UAV, on the same area, but without determining control points on the ground. The flight was made during the 27 November 2018period. The data were processed with the same type of program and we obtained a new digital terrain model, for the same area [11].

Verification of results

During the systematic cadastral work, we performed measurements with the total station and GNSS receivers to determine the property limits.To do this, we created a topographic network of densification, determined with the Leica GS08 Plus GNSS receivers connected to the National Network of Permanent GNSS Stations (RN-SGP) through the ROMPOS system. The network was made of wooden stakes.

From this network were selected a number of 31 points spread over the entire area over which the flight was made and the digital terrain model was obtained (Figure 4).The accuracy of determining these control points is ± 3 centimeters. The points were named: B01, B02, B03, B04, B05, B06, B07, B08, B09, B10, B11, B12, B15, B16, B19, B21, B23, B38, B43, B50, B54, B55, B65, B68, 5, 6, 7, 8, 9 10 and B64.

Figure 4. Position of the control points.

The altitudes of these points were noted in a (Table 1), together with the name of each. The first digital model was loaded in the Agisoft program and, based on the planimetric position of each of the 28 points it was extracted from the digital terrain model, the altitude of each point and was noted in the table.

Table 1: Measurement results.

Similarly, the altitude of each of the 31 points of the second digital terrain model was determined and noted in the table, in the respective column.

Results and Discussion

Following the results obtained in Table 1 was made, in which we have:

In the column marked with 1, the altitudes of the 31 control points determined on the ground with GNSS technology.

In the column marked with 2, the altitudes of the 31 control points extracted from the digital terrain model obtained with the UAVDroneZone XF8-CT.

In the column marked with 3, the altitudes of the 31 control points extracted from the digital terrain model obtained with UAVSenseFlyeBee X.

The following columns are the differences between the altitudes of the control points determined in the 3 variants.

Thus, column 4 represents the difference between the altitudes obtained from GNSS measurements and the digital model obtained with the UAVDroneZone XF8-CT. Column 5 represents the difference between the altitudesobtained from GNSS measurements and the digital model obtained with UAVSenseFlyeBee X. Column 6 represents the difference between the altitudes obtained from the digital model obtained with the UAV DroneZone XF8-CT and the digital model obtained with the UAV SenseFlyeBee X in Figure 5.

Figure 5. Placement of the control points B.23 and B.55.

A graph of the altitude differences obtained in the 3 variants was made (Graph 1).

Graph 1. Differences in altitude.

Conclusions

We start from the hypothesis that the correct altitudes are those determined directly on the ground, with the GNSS technology, using the permanent stations of the National Agency for Cadastre and Real Estate Advertising. From the presented values, the altitudes extracted from the model made with DroneZone XF8-CT are very close to those determined on the ground.This is because 6 ground control points were used, which fixed the digital model. The altitude values determined with SenseFlyeBee X are about 31 centimeters higher on average.

Between the GNSS measurements and the digital model obtained from the DroneZone XF8-CT flight, the largest negative difference is -14 centimeters and the largest positive is +10 centimeters. These differences are due to the fact that points B23 and B55 are located in grassy areas, where the altitude given by the drone is not very correct because it stops at grass level and not at ground level. From Figure 3 it can be seen that the two points are not located on a flat area, such as an asphalt road, a concrete platform, etc. The mean difference between the two altitudes is -5.9 centimeters and falls within the accuracy of the GNSS determined point network, of ± 3 centimeters.

Figure 3. The Digital Terrain Model obtained from the cloud of points processing. The flight was realized in the days: 06.02 -07.02.2018.

The digital model made from the DroneZone XF8-CT flight is different on average against the topographic network with -31.7 centimeters. It is clear that the relatively constant difference comes from the fact that on this flight we had no control point, so that the digital model is higher than the real model, determined by measurements referring to a system of verified altitudes. The conclusion is that even if we use a UAV that has a powerful GNSS receiver, a few control points are however needed in order to have the absolute altitude of the digital model as close to the reality on the ground.

Significance of the Work

Currently, the UAV technology for the realization of topographic plans is increasingly used. UAVs are equipped with GNSS receivers that give the position of the points with very high accuracy. Often users prefer not to determine control points anymore, considering that the results obtained are correct.

The present paper demonstrates that, in order to have a correct topographic plan, close to the reality on ground, at a UAV flight, control points measured on the ground are required.

REFERENCES

1. PÃÂ??unescu C, Nache F, PÃÂ??unescu V. Reliable 2D crowdsourced cadastral surveys: case studies from Greece and Romania. Inter J Geo Info. 2019;9(2):89.
2. Paunescu C, Dimitriu GS, Mocanu V. Position determination using GNSS. Inter J Geo Info. 2013;16(7):205.
3. Torhonen M, Salzmann M, Kasperavicius R, Sas E, Hokkanen J. Sustainabl governance of european cadastre and land registry agencies. Proc World Bank Conf. 2015;11:23-27.
4. Potsiou C, Volakakis M, Doublidis P. Hellenic cadastre: State of the art experience, proposals and future strategies. Comput Environ Urban Syst. 2001;25:445-476.
5. Amhar F, Jansa J, Ries C. The generation of the true orthophotos using a 3d building model in conjunction with a conventional DTM. Int Arch Photogramm Remote Sens. 1998;32:16-22.
6. Ambani S, Kalinga J, Lemmen C. Handheld land administration mapping methods: Institution of surveyors of kenya demonstrates alternative approaches. GIM Int. 2017;31:25-27.
7. Lengoiboni M, Lemmen C, Zevenbergen J. Innovative tools for tenure recordation: Documentation of overlapping and secondary land rights. GIM Int. 2017;31:41-43.
8. Williamson I, Rajabifard A, Kalantari M, Wallace J. A land information: Accurate, assured and authoritative. Proc FIG Reg Conf. 2012;8:26-29.
9. Basiouka S, Potsiou C. The volunteered geographic information in cadastre: Perspectives and citizens’ motivations over potential participation in mapping. Geo J. 2014;79:343-355.
10. Basiouka S, Potsiou C. A proposed crowdsourcing cadastral model: Taking advantage of previousexperience and innovative techniques. Surv Rev. 2016;12:419-433.
11. Mourafetis G, Apostolopoulos K, Potsiou C, Ioannidis C. Enhancing cadastral surveys by facilitating theparticipation of owners. Surv Rev. 2015;47:316-324.

Citation: Paunescu C, Marinescu CM, Paunescu V (2020) Comparative Study on Executing Topographic Plans Using UAVs. J Remote Sens GIS. 9:279. DOI: 10.35248/2469-4134.20.9.27

Copyright: © 2020 Paunescu C, et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.