Data Acquisition Tree Height and Areal Parameters of Mangrove Vegetation Using UAV

. Unmanned aerial vehicles (UAVs) can be used to assess the biophysical parameters of mangroves (particularly tree height and area). UAVs can produce high-resolution imagery, in this study UAV imagery was used to calculate tree height distribution and mangrove area. UAV imagery was taken using a commercial UAV with a planned flight path. The obtained UAV images are then compiled based on coordinate information and z-elevation values, and then 3D modelling is created from the point cloud. The 3D model will be detailed by a dense point cloud. The surface area and texture are obtained from the 3D mesh. Finally, a high-resolution DSM can be generated. DSM is the elevation of the overall surface including vegetation, buildings, and others, while DTM is the elevation data of the ground surface, so the next process is to separate bare ground and above-ground objects. Tree height is obtained from the intersection of DSM and DTM elevation data. The area of mangrove vegetation was obtained using the on-screen digitisation method. Based on the results of the analysis conducted by comparing tree height data from UAV images and in-situ measurement tree height data has a difference with a range of 0.24 m to 0.94 m, with an average of 0.49m. This remote sensing survey research using commercial UAVs can be concluded as one of the solutions to acquire the parameters of tree area and height distribution with good accuracy.


Introduction
Remote sensing is one of the alternative methods to collect information in effective and efficient ways.Remote sensing is a science or method of collecting information about natural phenomena on objects without direct contact with these objects.Remote sensing is one of the technologies widely used in marine research.One of the remote sensing tools in the marine field is satellite 1 .The results of processing satellite image data using software produce accurate and fast data.There are several technical weaknesses in obtaining satellite image data, including frequent interference from clouds so that clouds cover the resulting image.Another weakness is the location of the satellite far from the earth's surface so that the resolution of the images obtained is not too high.To overcome some of these weaknesses, a new alternative is now emerging, namely the use of UAVs to generate image data.UAVs can fly low below the clouds, so the resulting data is high-resolution and undisturbed by clouds.Another advantage is that drone images can be used at flexible times and locations 2 Mangrove forests are very important 3 and dynamic ecosystems that provide benefits to the environment.These benefits include blue carbon storage 4,5 , biodiversity support 6 , sustainable fisheries 7 , coastal defense 8,9 , and even tourism 10 .Unfortunately, mangroves face considerable threats 11,12 and have suffered considerable damage since the 1980s 13,14 .The Global Mangrove Alliance (GMA) has set important goals that aim to prevent mangrove forest shrinkage, expand restoration efforts, and raise awareness to improve mangrove forest conservation 15 .Monitoring and measuring the parameters of mangrove vegetation, such as tree height and mangrove area, is essential to obtain the right information for conservation, climate change mitigation, and disaster risk reduction in coastal areas.Conventional fieldbased data collection methods for mangrove ecosystems are often labor-intensive, timeconsuming, and not cost-effective 16 , especially when monitoring large or remote areas.Based on this, the integration of remote sensing technology has emerged as a very important approach to improve data acquisition in mangrove ecosystems 17 .UAVs are instruments with high spatial resolution 18 , which enables detailed images of vegetation structure and ground surface characteristics 19 .UAVs can conduct surveys on demand, making them suitable for small-scale monitoring of specific areas of need 20,21 .Several studies on mangrove tree height measurements using UAVs have been conducted [22][23][24][25] .some of them use high-resolution UAV imagery and a combination of LIDAR and other complements.this study is a bit different because we only use commercial UAVs with fairly good results.in this study, we aim to obtain the height distribution of mangrove trees and the area of the mangrove ecosystem.

Fig. 2. UAV flight planning
Protractor: In addition to remote sensing data, in-situ mangrove tree height measurements will be made using a protractor (trigonometric angle measurement method).This method involves measuring the angle from the observer's eye level to the tree top and then using trigonometric calculations to determine the tree height.Several in-situ measurements will be conducted at several locations within the study area to provide field validation data.The mangrove vegetation used is to take 6 samples of determination trees, the trees used as samples are trees that are easy to determine the height and besides that, the trees are easily identified when viewed on the image of the drone data.This is to avoid in-situ data and different drone image data.

DSM and DTM extraction
The captured data will then be processed using 2 software tools, namely 2.2.1.Agisoft Photoscan Professional which has a function to process the results of UAV photos taken into 3D, then it will be processed into Digital Surface Model (DSM) data which has a land surface elevation value.2.2.2.QGIS (SAGA) has a function to extract bare earth data and objects above ground.bare earth data will then be analyzed close gaps to interpolate DTM values and close empty values on the layer.

Area and Height Distribution Calculation
The area of mangrove vegetation will be calculated using the digitization process by applying the community unit technique.The technique is based on mangrove vegetation cover.Tree height will be calculated by subtracting the DSM height value from the DTM height value at the same location.This subtraction will generate mangrove tree canopy height values across the study area, and then input tree point vectors to calculate mangrove tree height distribution.

Validation and Accuracy
The height of the tree obtained is then compared with the measurement results using a protractor.Accuracy is calculated based on the elevation value of each measurement, the accuracy value is also known from the RMSE (Root Mean Square Error), MAD (Mean Absolute Deviation), and MAPE (Mean Absolute Precentage Error) values.

Results and Discussion
Aerial imagery data generated is data taken in the coastal area of Kedatim Village, Saronggi District, Sumenep Regency.Drone-generated images with a total 139 photos.The aerial photo data was then combined using Agisoft software.The resulting data is slightly blurred in some parts, this is expected due to the process of taking aerial photography data or external constraints (wind).The output of data processing is the area and height distribution of trees, the results are presented in the form of a map.The results of the mangrove area obtained are 77015 m 2 , obtained from the digitization process by applying the community unit technique.The technique is based on mangrove vegetation cover.The condition of mangrove ecosystem in this study has two types of conditions, namely the first is a tight mangrove community.The second is a mangrove ecosystem that is not too dense (there are surrounding water areas) and even tends to have more water components, so the digitization is only done on mangrove vegetation, and water areas are ignored.The advantage of using the community unit technique with the digitisation process is that it can obtain results that have a good level of accuracy 26 .The next process after digitizing is to calculate the mangrove area by utilizing the calculate geometry feature in QGIS.The results of the height distribution of mangrove trees were successfully identified with a total number of mangrove individuals of 6858 units.The accuracy of the number of individual mangrove trees using UAV imagery is 51.11% 27 .This value is not too high due to possible interference from devices and natural conditions.The height of the drone when taking aerial photos is very important to note.Strong winds will disturb the UAV in taking aerial photos and cause the UAV to move away from the flight path, resulting in a gap in the orthophoto results.To avoid gaps or displacement, the overlap used is more than 60%, set a flight height of 80-150 m to get good spatial resolution, and a flight speed of 10-15 m/s to avoid large Apparent Image Motion values 28 .A large AIM value can cause objects in the photo to become blurred and the orthophoto results become difficult to interpret because of the movement, so the recognition of the photo object becomes more difficult 29 .Measurement of mangrove tree height aims to determine the distribution of mangrove trees in the research location.The mangrove tree height measurement technique is based on the value of difference between the DEM (Digital Elevation Model) value and the DTM (Digital Terrain Model) value 30 , the DSM to DTM extract process uses the SAGA module in QGIS, the DSM to DTM extraction process has the aim of removing objects contained on the ground surface, and leaving only the height of the land surface.The difference between DSM and DTM data is calculated to determine the value of tree height.1-5 2250 4.
15-18 82 Total 6858 The height of mangrove trees in the study area is dominated by mangroves with a tree height of ˃10-15 m. mangrove tree height between 10 to 15 m is classified as a mature mangrove area 31 .Each type of mangrove has different characteristics, including mangrove tree height.These characteristics cause diversity in mangrove tree height.The accuracy of this estimation was tested by validating the calculated tree height (UAV) with the Tree height (Protractor) observed in the field (in-situ).The comparison results showed that the difference in mangrove height from UAV and in-situ was found to be about 0.24 to 0.94 m with an average of 0.49m.The MAD value is 0.486.The RMSE value is 0.7.The MAPE value was 7.8%.These errors when compared to the results of previous studies, indicate that the accuracy of tree height estimation calculated in this study is lower than previous studies that can show vertical accuracy with R square values of 0.9; 0.86; 0.76, for example, [32][33][34] .The high error introduced in this study may stem from the low accuracy of the handheld GPS used for mangrove sampling, leading to positional mismatches between the estimated tree height (UAV) and the tree height (protractor).However, this error is still much smaller than the RMSE obtained by Castellanos-Galindo et.al. 35, who calculated the height of mangrove trees on the northern coast of the Nariño Department of the Colombian Pacific coast, using a UAV.The R Square value they obtained was only 0.06.Then the R square value of 0.7 is still acceptable.Research using unmanned aircraft is recommended to obtain more effective and efficient data.This is because taking tree height using a protractor often has problems determining the highest tree shoots due to dense mangrove forests.The difficulty level of mangrove area data collection is also influenced by natural conditions and the density of the mangrove ecosystem.While data collection using drones is more measurable error value.In addition, in scientific publications regarding the height of mangroves in an area, often use the average tree height.But in fact researchers only take the height of trees that are easy for researchers to measure where researchers will more easily measure trees that are taller than researchers than the height of the trees that researchers take.

Conclusion
The results showed an area of mangrove vegetation with an area of 77015 m 2 dominated by mangrove trees with a height of 10 to 15 m.The height is included in the category of mature mangrove ecosystems.Remote sensing survey research using commercial UAVs can be concluded as one solution to obtain the parameters of the area and height distribution of mangrove trees with fairly good quality, with an RMSE value of 0.7. in the future these parameters can be used in biomass calculations, carbon stocks, replanting planning, monitoring, and other of necessity.

2. 1
Data collection UAV image and protactor data were taken at the location of fig.1.Data collection was carried out on August 7, 2023 in Kebun Dadap Timur Village, Saronggi District, Sumenep Regency.

Fig. 1 .
Fig. 1.Location of the study area.(a) Location of the study area marked by a point.(b) Unmanned aerial vehicle (UAV) image of the study area.Yellow dots represent measurement locations using a protractor.UAV Image Data: Data collection using UAVs begins with making a flight path plan, taking aerial photos and GPS loggers.Aerial photographs are taken at an altitude of 100 m with front and side overlap of 80%. in the resulting image planning will obtain 139 images.The UAV used is a dji phantom 4 pro weighing 1380 grams, with specifications namely; 20 MP camera, pix resolution 2.40 cm/pix, flight time up to 28 minutes, and has a maximum speed of 20 km/s.flight path and other information can be seen in fig. 2.

Fig. 3 .
Fig. 3. Areal Parameter of mangroves.Measuring mangrove area using the community unit technique with a digitization process on the screen.digitization process using QGIS software.

Table 1 .
Conferences https://doi.org/10.1051/bioconf/2024890700289 SRCM 2023 Class range of tree height values and number of individuals per class No

Table 2 .
Validation: (a) Comparison of observation measurement results (UAV) with