Application of Fixed-Wing UAVs to Develop Digital Terrain Model on Coastal Peatland Bengkalis Island

Destructive abrasion is currently occurring in Teluk Papal Village. The coast of Teluk Papal Village has a humus soil structure with soil conditions with a type of peat that is relatively easy to decompose. The cause is along the coast because of the absence of coastal protection plants (mangroves). To prevent the increasing abrasion rate, a comprehensive plan is needed by involving relevant stakeholders in both short and long-term planning. One of the efforts in controlling the abrasion rate is the construction of breakwater buildings. Before the construction is carried out, a study is needed to support the breakwater construction. One of the supports is to make a topographic or contour map. The field data acquisition process uses Unmanned Aerial Vehicle (UAV) technology. This data processing technique uses photogrammetric data processing techniques with the stages of flight path planning, measurement of Ground Control Point (GCP), and Benchmark points, which are then tied to BIG's CORS points to produce higher control point accuracy, Dense Cloud Point to create Digital Elevation Model (DEM) or Digital Terrain Model (DTM), and orthophoto embellishment. Furthermore, horizontal and vertical position accuracy tests were conducted using the calculation method issued by BIG regulation No. 15/2014 to provide information on how far the accuracy of the resulting map is. The results of the Digital Terrain Model (DTM) and orthophoto data accuracy test research based on the results of the accuracy test obtained a LE90 value of 0.6757 meters, then for the CE90 value of 0.1543 meters so that it meets the map requirements at a scale of 1:2500 at class 1 horizontal accuracy and class 2 vertical accuracy.


Introduction
Teluk Papal is one of the villages in the Bantan sub-district, a settlement with most of the population as fishermen. The coast of Teluk Papal Village has a humus soil structure with soil conditions that are a type of peat that is relatively easy to decompose. Lack of coastal protection plants (mangroves) as a binder, causing a severe abrasion problem along the coastline of Teluk Papal Village.
According to Sutikno (2014), the coast of Bengkalis Island, which experiences abrasion per year, has increased with an average of 59.02 ha/year for the last 26 years. The Central Government, through the National Development Planning, designated Bengkalis Regency, Riau Province, as a priority National Strategic Activity Center (PKSN) in the 2020-2024 National Medium-Term Development Plan (RPJMN). This is a big hope for the local government to bring Bengkalis even further forward. To realize this, support from the Riau Provincial Government is needed to accelerate ministries and institutions together.
Development requires a plan, be it long-term, medium-term, or short-term. Infrastructure development planning and countermeasures in abrasion-prone areas need to be supported by several data. One supporting data required is the abrasion map of Teluk Papal Village. Therefore, it is necessary to do topographic mapping and its appearance. The map information can support further planning for developing and preventing abrasion vulnerability in Teluk Papal Village. The process of field data acquisition for making abrasion maps directly in the field with an area of ± 400 hectares will undoubtedly require a long time, a considerable cost, and much personnel. However, the development of mapping technology will undoubtedly facilitate the field data acquisition process, and various field data acquisition methods are currently developing. With the guidance to produce fast, inexpensive map data with few personnel and the maps created can be accounted for, a suitable field data acquisition method is needed.
One is the field data acquisition method using the Unmanned Aerial Vehicle (UAV) aerial photo shooting method. Shooting with aerial photography is favored because of its low cost, few personnel, and fast field data acquisition.
Unmanned aircraft technology, which is now gaining popularity, often referred to as Unmanned Aerial Vehicles (UAVs) or drones, is one of the new alternative technologies for mapping, especially aerial photography. Foreign and domestic researchers and practitioners increasingly use and develop UAVs for mapping applications. UAVs have become an alternative to cheap remote sensing technology as a source of spatial data (Bendea, H et al., 2008;Adam S.M et al., 2011;Rokhmana, 2012).
This research aims to conduct topographic mapping of abrasion disaster in the Desa Teluk Papal area, Bantan Bengkalis District, Riau Province, by utilizing UAV shooting data to produce elevation data (DTM / DEM) that is accurate and has high accuracy. The results of this study are expected to be used as information in making efforts to reduce the risk of disasters that may occur for interested parties and the government and as a reference for rehabilitating the environment for the sustainability of the coastal area of Teluk Papal Village.

Data and Methodology
2.1. Aerial Photography Wolf (1993) Aerial photographs are photographs obtained from aerial surveys, namely shooting through the air in certain areas with specific photogrammetric rules. The result is a detailed record of the earth's surface, influenced by several factors, such as the camera lens's focal length, the aircraft's altitude, and the shooting time. Aerial photographs are further classified as vertical aerial photographs and inclined aerial photographs. Vertical aerial photography is when the camera axis at the time of shooting is entirely vertical or slightly tilted by no more than three. Most of the aerial photographs belong to the type of vertical aerial photographs. The second type of aerial photograph is a skewed aerial photograph, which is when the axis of the photograph is tilted between 3° and 90° from the vertical position. If the horizon is not visible, it is called inclined. If the horizon is visible, it is called highly inclined. Then in more detail, according to Wolf (1993), aerial camera orientation is divided into three: vertical, slightly inclined, and highly inclined.

Aerial Photography
Unmanned Aerial Vehicle (UAV), or Unmanned Aerial Vehicle (PUNA), is an aircraft that can operate without a crew. UAVs operate with an aircraft control operator outside the aircraft. In contrast, the aircraft operates automatically according to the command of the control operator, who is a visualization of the UAV aircraft.
Historically, UAVs were used by the military to practice shooting, where UAVs became the object of firing targets. Until now, the military world is still using UAVs as military tools. Besides the military world, in the private and civilian world, UAVs are also widely used for image capture tools because UAVs have embedded cameras. The general function is as a documentation tool for both photos and videos. However, its particular function is as a mapping survey tool. Aerial photography surveys using UAVs are much cheaper than aerial photography surveys using manned aircraft. Then to monitor conflict areas is also safer because there are no human operators on the aircraft. The flying ability of the UAV itself varies depending on the specifications of the UAV itself, but the recommended flying altitude is no more than 120 meters. At the same time, the maximum flying altitude, according to applicable law, is a flying altitude of no more than 150 meters.

Global Positioning System (GPS)
According to (Abidin, 2011), GPS is a radio navigation and positioning system using satellites owned and managed by the United States. Because it is based on satellites, GPS receiver observations do not have to be visible to each other (intervisibility), the important thing is that the receiver can capture GPS signals (satellite visibility) until it is sufficient. In principle, the greater the number of satellites that can be captured, the better the quality of GPS position acquisition. By utilizing the signal from the satellite then, the GPS receiver performs a backward binding to the specified base point (differential method). Then in terms of position observation using GPS, at least two GPS receivers are used. Then in recording data, the receiver must be in a static or stationary state because the observations are observations for one point. In GPS observations, the mandatory components include GPS receivers, sufficient satellites, and users.

Ground Control Points (GCP)
Ground control points are objects on the earth's surface that can be identified and have spatial information by the mapping reference system. Spatial information in the form of X, Y, Z, or Latitude-Longitude coordinates and the height of each GCP is measured using a sub-meter accuracy geodetic GPS. The essential need for GCP is the georeferencing process of photo processing results so that it has a reference system for what is needed in the mapping results. GCP is also used during data processing to help the geometric correction process on the orthophoto mosaic so that the accuracy of the resulting map will be high. Specifically, GCP also functions as: a.
The determining factor of geometric accuracy of photo processing results (orthophoto, DSM, DTM), the more thorough the GCP, the better the geometric accuracy of the output (with the rules of GCP placement being met). b.
Factors that facilitate the process of relative orientation between photos so that the presence of GCPs can improve the geometric accuracy of photo maps. c. A correction factor for photo processing results in a ball effect or error that causes the 3D model to be convex in the center of the measured area. d.
Factors that make it easier to unify the results of separate data processing, for example, processing data for areas A and B more quickly and effectively, rather than the unification process based on the entire point cloud (millions of them), which will take much time.

Fig 2. GCPs of the Research Site Plan
The use of GCPs is optional. GCPs help to increase the accuracy of the resulting map (up to ± 10 cm), so the consequence of not using GCPs is only that the accuracy of the resulting map becomes low (between ± 6 -12 m). The use of GCPs is also regulated according to the Indonesian National Standard (SNI), which is a maximum distance between GCPs of 2.5 kilometers. GCP installation takes quite a long time, with a capacity of 6-10 GCPs/day (according to field conditions), which is done before the aerial photography data acquisition process. For the case of making topographic maps, the role of GCP is quite essential. By using GCPs, the resulting topographic map can have a high Z accuracy so that the geographical conditions in the area can be analyzed with high confidence. Each GCP must have a premark or mark to be visible on aerial photographs. The premark can be a circle or a cross (+) that has 4 wings and intersects the control point. The premark that will be installed is an orange fabric mark, with the minimum size of the premark on aerial photos being 10 pixels long and 3 pixels wide for each premark. Each wing of the premark. The actual premark size in the field adjusts the ground resolution value of the aerial photography or about 100 x 40 cm. The coordinates of the control points will be measured using Geodetic GPS with the RTK system. RTK (Real-Time Kinematic) system is a commonly used acronym for differential real-time positioning systems using phase data. To realize the real-time demands, the reference station must transmit phase and pseudorange data to the user in real-time using a specific data communication system. All GCPs are tied to one Benchmark of the Geospatial Information Agency (BIG) located around the area as a local base. Using this method, the resulting map will conform to mapping standards and have a global coordinate reference.

Geometry Accuracy Test
Geometry Accuracy is a value that describes uncertainty in the position coordinates of an object on the map compared to the position coordinates object that is assumed to be its actual position. The geometric accuracy component consists of Horizontal accuracy and vertical accuracy. The position accuracy test was carried out to get the map's confidence level of 90% Circular Error and Linear Error. Test Positional accuracy is determined using satisfactory test points provisions of the object used as a test point, namely: 1. Can be identified in the field and on the map to be tested. 2. It is an object that remains relatively unchanged over a period of time short time. 3. Have an even distribution throughout the area to be tested. Positional accuracy testing refers to the difference in coordinates (X,Y,Z) between points test on an image or map with the actual location of the test point on ground level. Accuracy measurement using root mean square error (RMSE) or circular error. On the necessary two-dimensional mapping taken into account are the (X,Y) coordinates of the test point and the actual position at the field. Analysis of positional accuracy using root mean square error (RMSE), which describes the value of the difference between the test point and the actual point. RMSE is used to describe accuracy, including random error, and systematically (Head of BIG No. 15, 2014).

Basic Map Geometry Accuracy
The accuracy of the orthophoto map is tested by following the accuracy of the Fine Map Bumi Indonesia (RBI) issued through the Regulation of the Head of the Agency Geospatial Information (BIG) No 15 of 2014. According to the Regulation of the Head of BIG No. 15 of 2014, the geometric accuracy of the map is obtained based on the provisions as follows: Will qualify if Circular Error (CE90) and Linear Error (LE90) get a confidence level of 90% or not more than the accuracy that has been determined according to the scale of the resulting image.

Tools and Materials
Tools and materials used in this research:

Results and Discussion
This section explains in detail the research conducted, including the research location, the data used, and the methodology applied.

Photogrammetric Data
Photogrammetric data acquisition using UAV Fixed wings produced a total of 536 photos. Of the total photos, several photos were not optimal when acquiring aerial photo data. Therefore, researchers checked the entire data and selected unusable photo data. The results of the data elimination process were obtained from as many as 534 photos, which were used for processing. Acquisition of ground control points (GCP), The process of acquiring ground control points requires an observation time of ± 10 minutes for one observation point. GCP result data is obtained in Table 1.  GCP 3 198657.138 171686.190 4.136 GCP 4 199041.553 171785.883 4.599 GCP 5 199409.304 171704.830 3.850 GCP 6 196994.902 171982.488 3.991 Table 2 is the value of Ground Control Point data acquisition in the coastal land area of Teluk Papal Village, which is used to correct geometry and elevation data in photogrammetry to produce high orthophoto accuracy data. The observation method used in this research is the Real Time Kinematic (RTK) observation of the Indonesia Continuously Operating Reference Station (Ina-Cors) system. To find out the distribution of GCP control points can be seen in Figure 4.

Data Processing
Data processing is carried out to combine aerial photography data with ground control point (GCP) data to provide detailed information on the topography of the Teluk Papal Village Coast. At this stage, the photo data is processed using Agisoft Metashape software. The steps of processing aerial photography data using Agisoft Metashape software are described as follows: 1. Add Photos 2.
Align Photos 3. Import GCP Coordinates and Identify GCP Points 4.
Build Dense Clouds Point 5.
Build Digital Elevation Model (DEM) Digital Elevation Model is a digital terrain model in raster or grid format. From DEM data, elevation information such as cut and fill can be derived for other modeling. There are two terminologies related to DEM: DSM (Digital Surface Model) and DTM (Digital Terrain Model). The results of survey data obtained from processing using Agisoft Software are as follows : • A total of 536 photos are included, after filtering, a total of 534 photos were obtained.

•
The area is 4.31 Km 2 .

•
The total Tie Points are 522,693.

Ground Sampling Distance (GSD) Result
GSD or ground sampling distance measures the pixel resolution of the results in aerial photographs, both aerial photographs with a metric camera and aerial photographs with a non metric camera. The GSD value is often used as a benchmark for quality resulting aerial photographs. By setting the GSD value, the user requires flying height and camera resolution with clear parameters. Determine the GSD value using the following formula : The following is a valid table of shooting results in the field using Sonny A6000 series cameras processed with Agisoft Software : Table 3. Camera Report With the formula above, you can find the value.
= 320 20 × 4.19 = 6.701 / From the above results, it can be seen that the difference from the ground sampling distance value (GSD) the calculation results obtained are 6.701 cm/pix, while the values in the report agisoft software obtained are 5.72 cm/pix. These results have a difference of 0.981 cm/pix.

Horizontal Accuracy
Position accuracy testing refers to the difference in coordinates (X, Y) between the test points on the map and the actual test points on the ground surface. Accuracy measurement uses root mean square error (RMSE). RMSE used to describe accuracy includes random error, and systematically, this RMSE can be calculated when the coordinate transformation is complete with the formula: 90 = 1.5175 × (2) ( ) = √ ((2 * (( ) ) 2 ) (3)

Table 4. RMSEr and CE90 Calculation Results
Based on table 4 shows the total horizontal RMSEr (X,Y) value worth 0.1017 m. The orthophoto accuracy value is the value (Circular Error) CE90 of 0.1543 for horizontal accuracy, which means that the position error orthophoto does not exceed the accuracy value with a 90% confidence level obtained by using the standard geometric accuracy in table 1.

Vertical Accuracy
RMSEz calculation results obtained from the calculation of the difference in value Z field and Z on DTM is 0.4095 m and the value is LE90 in this study is equal to 0.6757m. The DEM accuracy value is (Linear Error) LE90 for vertical accuracy, which means that the vertical error does not exceed the accuracy value with a confidence level of 90%. This RMSE can be calculated when the coordinate transformation is complete with the formula: 90 = 1.6499 × (4) The accuracy of the map is based on the regulation of the Head of the Geospatial Information Agency (BIG), each map generated using aerial photo data has map class accuracy value. The accuracy of the map class can be known by calculates CE90 and LE90 values, CE90 values for horizontal precision and LE90 for vertical accuracy. After knowing the value of CE90 and LE90 Of course, it can be done by dividing the map accuracy class at a specific scale. The accuracy of the 1:2500 Scale Map is as follows: It can be concluded in table 6 that the results of the CE90 and LE90 tests are known included in the map scale of 1:2500 with horizontal accuracy of class 1 and vertical accuracy in class 2.

Orthographic Map of Abrasion in Teluk Papal Village
Orthophoto maps are the results obtained from shooting aerial photographs using unmanned vehicles. Orthophoto maps are produced after several processes, including in-camera orientation correction, out-of-camera orientation correction, transforming model coordinates into ground coordinates, and making aerial photo mosaics. Making map layouts using ArcGIS 10.7 Software on the Universal Transverse Mercator Zone 48North (UTM 48N) coordinate system: map information, grid, map source, and other information as in making maps in general. Topographic map results are obtained by processing from the Agisoft Metashape application to produce Digital Terrain Model (DTM) data. The process carried out to produce DTM data is the result of Ground classification and extracting areas that contain information, such as buildings and trees around aerial photographs, after generating data to produce a Digital Terrain Model (DTM) in the form of information on the state of land contours in the abrasion area of Teluk Papal Village.  Topographic maps classified with Agisoft Metashape Profesional software can be displayed as ten contour profiles in the land area towards the sea. Teluk Papal Village.

Fig 8. Profile division
The profile data shows that the Digital Terrain Model (DTM) data generated from the extraction process has data quality close to the land's actual topography.

Fig 9.
Cross Section of profile 2 3.11. DSM Export Results In processing aerial photos using Agisoft Metashape Professional software, geometry formation is also carried out, which will produce DSM. The following will show the results of the build geometry that produces DSM.  Figures 9 and 10 show that the Digital Surface Model (DSM) data generated from the extraction process has data quality that is close to the topography of the land and the projections on the actual land.

3D Contour of Abrasion of Teluk Papal Village
In obtaining 3D contour data, the data that must be prepared is contour data with elevation values with DTM data. Furthermore, the two data will be overlaid using arc scene software to generate the appearance of a 3D map.

Conclusions and Suggestions
After conducting the research, there are conclusions from the author as follows: 1. Unmanned Aerial Vehicle (UAV) technology can be utilized to map an area of ± 400 hectares with a single flight of ± 350 hectares, depending on the battery capacity and size of the UAV. Data acquisition is carried out for four days, so using this technology will save time, energy, and costs.

2.
Digital photogrammetric processing utilizing 6 GCPs can be optimized for the study area with an area of ± 400 hectares and on a notso-flat land contour area. This is evident from evaluating the accuracy of photogrammetric processing maps using Agisoft Metashape Software according to