*2.1.1 In environmental sciences*

In environmental sciences, the applications of these three techniques are very diverse and operate at different spatial scales.

*Spatial Variability in Environmental Science - Patterns, Processes, and Analyses*

large amount of elevation data that are transformed into DTM are used.

Surface Model (DSM).

dimensions.

The third category encompasses the methods of photogrammetric processing that allow to extract elevation data from aerial photographs; this technique also allows to collect a very large amount of data that are transformed into Digital

The evolution of these techniques can also be conceptualized in a "measurement method paradigm" (**Figure 1**): with the methods of the first category, we start from the images to make measurements. Paper images (maps or aerial photographs) available in paper format are scanned or acquired directly in digital format; geometric deformations are corrected, and they are georeferenced; measurements are then carried out and statistically processed. It should be noted here that the analysis is in two dimensions (in plan). With the methods of the second category, measurements are performed and transformed into images (DTM and DSM), which are the result of calculations and not a direct capture of the reality of the terrain. One of the major interests of these methods is to be able to work in three dimensions: the measurements made are in plan and in altitude. The third category closes the paradigm: the work process starts with images (aerial photographs), continues with measurements, and returns to images (DSM). Here, the analysis is also in three

This paradigm has undergone a twofold evolution over the last 25 years or so. First, the transition from 2D to 3D measurements has been a determining factor in coastal geomorphology. This has made it possible to objectify the position of the coastline, which is not always easy to detect in 2D aerial photographs [8]. This also made it possible to calculate volumetric and not just planimetric changes in beaches and dunes. In addition, with LiDAR technology and photogrammetry, there has been a progress toward higher density and accuracy of measurement and results. The first part of this chapter is devoted to a technical and conceptual synthesis. A bibliometric analysis shows the recent, rapid, and phased success of the use of GNSS, LiDAR, and then airborne photogrammetry in the field of coastal geomorphology. We will briefly describe the techniques, emphasizing their complementarities and their spatial and temporal scales of application. The second part of this chapter is an illustration of the first by selected examples of results obtained on the coast of northern France. The conclusion proposes a synthesis of the advantages and disadvantages of these techniques and discusses some future prospects.

photographs since the mid-twentieth century, and more recent high spatial resolution satellite images are often used. On historic maps that are often imprecise in their drawing, it usually results in approximate measurements. However, they represent valuable documents for a qualitative or naturalistic geo-historic analysis of landscape [3, 4]. On these maps published since the end of the nineteenth century and on vertical aerial photographs available since the 1930s, the successive positions of the coastline can be identified, digitized, and then compared in a Geographical Information Systems (e.g., [5–7]) in order to determine erosion or accretion zones. Within the second category of methods, in the shorter term, over the last 30 years or so, measurements have been carried out in two ways with high-tech instruments. First, geodetic instruments (e.g., total electronic stations and GNSS) are used to carry out data collection in the field. These measurements are timeconsuming and difficult to carry out over large areas; only a fairly limited number of points can be acquired (a few hundred to a few thousand); therefore, only beach profiles (cross section of the range from a few tens to hundreds of meters long) or Digital Terrain Models (DTM) of limited size (a few thousand square meters) and low spatial resolution can be surveyed. Second, also in the short term, telemetry instruments (e.g., ground or airborne LiDAR) that enable the rapid collection of a

**92**

Differential GNSS is an in-situ measurement system that is used in all areas where an accurate position on the planet's surface is required [9]. With GNSS, for example, time series of measurements can be made at fixed locations, for example, to identify the role of vertical movements of the continent in changes in relative sea level [10] or to map the plate boundary faults [11]. Multiple point measurements are also frequently taken at different locations and interpolated to produce elevation maps, for example, for glacial geomorphological mapping [12].

Light Detection and Ranging (LiDAR) is a telemetry system (remote measurement) that can be used in two main configurations: on the ground with TLS systems and in airborne configuration. As shown in a bibliometric study in the field of earth sciences in Ref. [13], the Terrestrial Laser Scanning (TLS) technique is used to make precise topographic surveys at distance without having to come into contact with the ground, which is very convenient in hard-to-reach terrain such as steep cliffs. Very large numbers of measurements can be made with these instruments, which allow to model the investigated object. Fields of investigation using the techniques are, for example, at fine-scale mineralogy and petrology and at a larger scale structural geology, seismology, volcanology, tsunami hazards, geomorphology, and cryosphere studies [13]. Airborne LiDAR is used to produce accurate elevation maps at a larger scale, usually over large areas. The spectrum of applications in the Earth and ecological sciences is wide [14], for example, for lava flow survey [15] and forestry [16, 17].

Photogrammetry can also be implemented at different scales, from very fine scale, over a few millimeters, for example, to monitor the rock surface weathering [18], to medium scale, over areas ranging from a few thousand square meters to a few hectares, for example, to monitor the evolution of slopes [19, 20] or coastal habitats [21–23]. Finally, it can be deployed on a large scale, over areas of several square kilometers (e.g., geomorphological mapping in high mountain environment) [24].

#### *2.1.2 In coastal geomorphology over the past 25 years*

In order to estimate the development of these three techniques in coastal geomorphology in particular, a bibliometric analysis was conducted following a method similar to that of [13] for the use of TLS in Earth sciences and [25] for UAVs in agriculture and forestry.

The Scopus database was used for finding publications (articles and book chapters) in the field of Earth and planetary sciences, from 1995 to 2019, searching for the following keywords: the kind of instrument (gps or dgps or gnss or dgnss/ LiDAR/uav or uas or drone), coast or beach or dune or shoreline. Early results often contained intruders (e.g., in the military, maritime navigation, biology, mathematics, or electronic fields). It was therefore necessary to reduce the research panel by excluding keywords and certain journals outside the field of coastal geomorphology.

**Figure 2** shows the results of this analysis for the three types of techniques. It can be seen that GNSS was used earlier than the other two techniques. This is logical since it is the first of the three technologies that have been developed. However, during the last years, the number of mentions of satellite positioning systems decreases. This is probably not because they are less used, rather because they have become commonly used and authors probably no longer feel the need to describe or even mention the technique in the publication. However, these GNSSs are still needed for complementing LiDARs and UAVs.

The LiDAR citation curve reflects the development of this technology during the last decades. In the second half of the 1990s, there are a limited number of publications in which the tool is experimental or is used by only a few organizations that

**95**

*Recent Advances in Coastal Survey Techniques: From GNSS to LiDAR and Digital…*

have the technical and financial means to use it (e.g., US Army Corps of Engineers USACE) [26]. This technique significantly increased at the beginning of the 2000s whether airborne or TLS (e.g., for monitoring the evolution of cliffs) [27–29]. The use of UAVs did not begin until the early 2000s. This technique took off during the second decade of the twenty-first century with the availability of light aircraft (<2 kg), ready to fly, simple to program and pilot, and financially affordable [22]. UAVs are most often used for photogrammetry (see below) and more rarely to carry small LiDAR systems [30] or multispectral sensors [31]. It is interesting to note that, at the end of the study period, the three techniques are almost equal in terms of citations in the literature: although the use of UAVs started late, it is now as common as

*Results of the bibliometric analysis on GNSS, LiDAR, and UAV mentions from 1995 to 2019.*

Global Navigation Satellite Systems (GNSSs), more commonly known by the acronym GPS, make it possible to measure a position in three dimensions, anywhere on the Earth's surface, using the trilateration principle (with three satellites) [32], that is, by referring to the very precisely known positions of the satellites in their orbits. The use of a radio beacon precisely geo-referenced on land, installed near the study site and communicating with the GNSS receiver via UHF waves, makes it possible to reduce the error margin of measurement, which ranges from several millimeters to centimeters, or even millimeters both horizontally and vertically. During fieldwork campaigns, when numerous measurements for generating accurate Digital Terrain Model are required, the RTK (Real Time Kinematic) mode is generally used: the GNSS automatically records elevation data points during the operator's movements following a previously defined time step or distance.

Elevation point measurements can be acquired on foot when the areas to be covered are fairly small (a few thousand square meters or even a few hectares) or with an all-terrain vehicle (e.g., a quad where the GNSS is mounted coupled to an inertial station to compensate for the vehicle's movements) driving slowly (between 10 and 20 km/h) for larger and fairly flat areas. Although larger areas can be covered with a vehicle, the density of points collected is rather low. In the example shown in **Figure 3**, the elevation data points were collected using a Leica 1200 GNSS, which

**2.2 The use of these techniques in coastal geomorphology**

*DOI: http://dx.doi.org/10.5772/intechopen.91964*

that of GNSS and LiDAR.

**Figure 2.**

*2.2.1 Differential GNSS*

*Recent Advances in Coastal Survey Techniques: From GNSS to LiDAR and Digital… DOI: http://dx.doi.org/10.5772/intechopen.91964*

#### **Figure 2.**

*Spatial Variability in Environmental Science - Patterns, Processes, and Analyses*

maps, for example, for glacial geomorphological mapping [12].

*2.1.2 In coastal geomorphology over the past 25 years*

needed for complementing LiDARs and UAVs.

Differential GNSS is an in-situ measurement system that is used in all areas where an accurate position on the planet's surface is required [9]. With GNSS, for example, time series of measurements can be made at fixed locations, for example, to identify the role of vertical movements of the continent in changes in relative sea level [10] or to map the plate boundary faults [11]. Multiple point measurements are also frequently taken at different locations and interpolated to produce elevation

Light Detection and Ranging (LiDAR) is a telemetry system (remote measurement) that can be used in two main configurations: on the ground with TLS systems and in airborne configuration. As shown in a bibliometric study in the field of earth sciences in Ref. [13], the Terrestrial Laser Scanning (TLS) technique is used to make precise topographic surveys at distance without having to come into contact with the ground, which is very convenient in hard-to-reach terrain such as steep cliffs. Very large numbers of measurements can be made with these instruments, which allow to model the investigated object. Fields of investigation using the techniques are, for example, at fine-scale mineralogy and petrology and at a larger scale structural geology, seismology, volcanology, tsunami hazards, geomorphology, and cryosphere studies [13]. Airborne LiDAR is used to produce accurate elevation maps at a larger scale, usually over large areas. The spectrum of applications in the Earth and ecological sciences is wide [14], for example, for lava flow survey [15] and

Photogrammetry can also be implemented at different scales, from very fine scale, over a few millimeters, for example, to monitor the rock surface weathering [18], to medium scale, over areas ranging from a few thousand square meters to a few hectares, for example, to monitor the evolution of slopes [19, 20] or coastal habitats [21–23]. Finally, it can be deployed on a large scale, over areas of several square kilometers (e.g., geomorphological mapping in high mountain environment) [24].

In order to estimate the development of these three techniques in coastal geomorphology in particular, a bibliometric analysis was conducted following a method similar to that of [13] for the use of TLS in Earth sciences and [25] for UAVs

The Scopus database was used for finding publications (articles and book chapters) in the field of Earth and planetary sciences, from 1995 to 2019, searching for the following keywords: the kind of instrument (gps or dgps or gnss or dgnss/ LiDAR/uav or uas or drone), coast or beach or dune or shoreline. Early results often contained intruders (e.g., in the military, maritime navigation, biology, mathematics, or electronic fields). It was therefore necessary to reduce the research panel by excluding keywords and certain journals outside the field of coastal

**Figure 2** shows the results of this analysis for the three types of techniques. It can be seen that GNSS was used earlier than the other two techniques. This is logical since it is the first of the three technologies that have been developed. However, during the last years, the number of mentions of satellite positioning systems decreases. This is probably not because they are less used, rather because they have become commonly used and authors probably no longer feel the need to describe or even mention the technique in the publication. However, these GNSSs are still

The LiDAR citation curve reflects the development of this technology during the last decades. In the second half of the 1990s, there are a limited number of publications in which the tool is experimental or is used by only a few organizations that

**94**

forestry [16, 17].

in agriculture and forestry.

geomorphology.

*Results of the bibliometric analysis on GNSS, LiDAR, and UAV mentions from 1995 to 2019.*

have the technical and financial means to use it (e.g., US Army Corps of Engineers USACE) [26]. This technique significantly increased at the beginning of the 2000s whether airborne or TLS (e.g., for monitoring the evolution of cliffs) [27–29]. The use of UAVs did not begin until the early 2000s. This technique took off during the second decade of the twenty-first century with the availability of light aircraft (<2 kg), ready to fly, simple to program and pilot, and financially affordable [22]. UAVs are most often used for photogrammetry (see below) and more rarely to carry small LiDAR systems [30] or multispectral sensors [31]. It is interesting to note that, at the end of the study period, the three techniques are almost equal in terms of citations in the literature: although the use of UAVs started late, it is now as common as that of GNSS and LiDAR.

#### **2.2 The use of these techniques in coastal geomorphology**

#### *2.2.1 Differential GNSS*

Global Navigation Satellite Systems (GNSSs), more commonly known by the acronym GPS, make it possible to measure a position in three dimensions, anywhere on the Earth's surface, using the trilateration principle (with three satellites) [32], that is, by referring to the very precisely known positions of the satellites in their orbits. The use of a radio beacon precisely geo-referenced on land, installed near the study site and communicating with the GNSS receiver via UHF waves, makes it possible to reduce the error margin of measurement, which ranges from several millimeters to centimeters, or even millimeters both horizontally and vertically.

During fieldwork campaigns, when numerous measurements for generating accurate Digital Terrain Model are required, the RTK (Real Time Kinematic) mode is generally used: the GNSS automatically records elevation data points during the operator's movements following a previously defined time step or distance.

Elevation point measurements can be acquired on foot when the areas to be covered are fairly small (a few thousand square meters or even a few hectares) or with an all-terrain vehicle (e.g., a quad where the GNSS is mounted coupled to an inertial station to compensate for the vehicle's movements) driving slowly (between 10 and 20 km/h) for larger and fairly flat areas. Although larger areas can be covered with a vehicle, the density of points collected is rather low. In the example shown in **Figure 3**, the elevation data points were collected using a Leica 1200 GNSS, which

**Figure 3.**

accuracy is ±1.5 cm horizontally and ±3 cm vertically, coupled to an inertial station SBG 500-E on a quad. The DTM was calculated by triangulation of Delaunay along the coastline of the port of Dunkirk. The measured points were superimposed on it. The intertidal zone was measured with a GNSS receiver of an all-terrain vehicle; the alignments of the points along the vehicle's trajectory are clearly visible. The upper beach, the foredune, the muddy areas, and the sand and pebble spit were measured on foot because the relief is too constraining for the vehicle; foot surveys enable higher measurement densities. This type of topographic survey is timeconsuming: for a survey of 36,360 points, over a surface area of 33.5 ha, resulting in a measurement density of 0.1 point per m<sup>2</sup> , it took 5 days of work during daytime and low tide periods.

#### *2.2.2 Airborne LiDAR*

Airborne LiDAR is an active telemetry instrument. A vector, aircraft [33–35], helicopter, or UAV [30, 36] whose position and altitude are precisely determined using a differential GNNS, carries a side-scanning laser that emits pulses under the vector. The laser beam is emitted toward the ground; the round-trip time of this beam and its "echo" is measured. The speed of propagation of the beam is perfectly known according to the atmospheric conditions (temperature and humidity). For each point aimed at on the ground, several signals are recorded and averaged: this is the distance between the device and the ground surface. Ground control points are checked with GNSS to calibrate the data as accurately as possible.

Airborne LiDAR surveys can provide a large number of measurements over large areas of several square kilometers. The measurement density is higher than that obtained with in situ GNNS. It can typically range from 1 to 2 points/m<sup>2</sup> , for example, from 1.2 to 1.4 points/m2 after data filtering [37, 38]. In the example shown in **Figure 4**, a Leica HawkEye III topo-bathymetric LiDAR sensor was used with a 500 KHz frequency in infrared spectrum for topographical surveying and a 35 KHz frequency in the green spectrum for bathymetric monitoring in shallow

**97**

density of 1–2 points/m2

**Figure 4.**

*Geosystems).*

*2.2.3 Airborne photogrammetry*

from different angles using the principle of stereoscopy.

*Recent Advances in Coastal Survey Techniques: From GNSS to LiDAR and Digital…*

water. The density of topographical measurements is 12 points/m2

filtering. According to the manufacturer, the measurement accuracy is ±20 cm horizontally and ±2 cm vertically. This vertical margin of error is generally noted on bare surfaces (sand and rock) but increases significantly in vegetated areas (25 cm). Maps produced from LiDAR data are Digital Surface Models (DSM) that show the top of all objects. These DSMs represent the ground relief in bare areas (beach and rocky plains) but may show the tops of vegetation or buildings elsewhere. In geomorphology, however, the object of the study is the relief on the ground and not that on the surface of the vegetation or other objects. To obtain a Digital Terrain Model (DTM), buildings can be removed by using a vector map of the building to delineate the areas of the points to be deleted. At the location of each building, an interpolation is then made with the ground elevations around the perimeter of the building. For vegetated areas, the data are filtered: the first echo of the signal is considered as the one corresponding to the surface of the objects, and the last echo (the furthest from the vector) is assimilated to the ground. DTMs are therefore calculated using the last echoes resulting in lower measurement densities. However, the vegetation is sometimes too dense for the LiDAR signal to reach the ground and for a DTM to be extracted from the DSM; this is notably the case in coastal dunes covered by sea buckthorn (**Figure 4**). DSMs or DTMs are calculated by interpolation from points; they can be represented in vector mode as contour maps or in raster mode where each pixel contains elevation information. With a data point

*Extract of an extensive 1 m spatial resolution DTM on the harbor of Dunkirk coast, northern France carried out with a Leica HawkEye III LiDAR (sources: map SHOM, ROLNP on data.shom.fr; picture, Leica* 

, there is no need to aim for a spatial resolution of DSM or

DTM finer than 1 m/pixel. **Figure 4** shows the very wide foreshore along the port of Dunkirk, which gradually narrows toward the east. It also shows a morphology of bars and troughs on the foreshore. The sand and gravel spits are clearly visible.

Photogrammetry is a nonactive telemetry process because no signal is emitted. It can be defined simply as the "science of making reliable measurements from photographs" [39]. It is most frequently applied to aerial photographs. It allows distance and surface measurements to be made in plan. It is also possible to extract altitude information from two vertical aerial photographs of the same area taken

before data

*DOI: http://dx.doi.org/10.5772/intechopen.91964*

*Recent Advances in Coastal Survey Techniques: From GNSS to LiDAR and Digital… DOI: http://dx.doi.org/10.5772/intechopen.91964*

#### **Figure 4.**

*Spatial Variability in Environmental Science - Patterns, Processes, and Analyses*

accuracy is ±1.5 cm horizontally and ±3 cm vertically, coupled to an inertial station SBG 500-E on a quad. The DTM was calculated by triangulation of Delaunay along the coastline of the port of Dunkirk. The measured points were superimposed on it. The intertidal zone was measured with a GNSS receiver of an all-terrain vehicle; the alignments of the points along the vehicle's trajectory are clearly visible. The upper beach, the foredune, the muddy areas, and the sand and pebble spit were measured on foot because the relief is too constraining for the vehicle; foot surveys enable higher measurement densities. This type of topographic survey is timeconsuming: for a survey of 36,360 points, over a surface area of 33.5 ha, resulting in

*In the field topographic survey using a GNSS; example of a DTM (the measurement points are overlapped on* 

Airborne LiDAR is an active telemetry instrument. A vector, aircraft [33–35], helicopter, or UAV [30, 36] whose position and altitude are precisely determined using a differential GNNS, carries a side-scanning laser that emits pulses under the vector. The laser beam is emitted toward the ground; the round-trip time of this beam and its "echo" is measured. The speed of propagation of the beam is perfectly known according to the atmospheric conditions (temperature and humidity). For each point aimed at on the ground, several signals are recorded and averaged: this is the distance between the device and the ground surface. Ground control points are

Airborne LiDAR surveys can provide a large number of measurements over large areas of several square kilometers. The measurement density is higher than that obtained with in situ GNNS. It can typically range from 1 to 2 points/m<sup>2</sup>

shown in **Figure 4**, a Leica HawkEye III topo-bathymetric LiDAR sensor was used with a 500 KHz frequency in infrared spectrum for topographical surveying and a 35 KHz frequency in the green spectrum for bathymetric monitoring in shallow

checked with GNSS to calibrate the data as accurately as possible.

, it took 5 days of work during daytime

after data filtering [37, 38]. In the example

, for

a measurement density of 0.1 point per m<sup>2</sup>

example, from 1.2 to 1.4 points/m2

and low tide periods.

**Figure 3.**

*the DTM).*

*2.2.2 Airborne LiDAR*

**96**

*Extract of an extensive 1 m spatial resolution DTM on the harbor of Dunkirk coast, northern France carried out with a Leica HawkEye III LiDAR (sources: map SHOM, ROLNP on data.shom.fr; picture, Leica Geosystems).*

water. The density of topographical measurements is 12 points/m2 before data filtering. According to the manufacturer, the measurement accuracy is ±20 cm horizontally and ±2 cm vertically. This vertical margin of error is generally noted on bare surfaces (sand and rock) but increases significantly in vegetated areas (25 cm).

Maps produced from LiDAR data are Digital Surface Models (DSM) that show the top of all objects. These DSMs represent the ground relief in bare areas (beach and rocky plains) but may show the tops of vegetation or buildings elsewhere. In geomorphology, however, the object of the study is the relief on the ground and not that on the surface of the vegetation or other objects. To obtain a Digital Terrain Model (DTM), buildings can be removed by using a vector map of the building to delineate the areas of the points to be deleted. At the location of each building, an interpolation is then made with the ground elevations around the perimeter of the building. For vegetated areas, the data are filtered: the first echo of the signal is considered as the one corresponding to the surface of the objects, and the last echo (the furthest from the vector) is assimilated to the ground. DTMs are therefore calculated using the last echoes resulting in lower measurement densities. However, the vegetation is sometimes too dense for the LiDAR signal to reach the ground and for a DTM to be extracted from the DSM; this is notably the case in coastal dunes covered by sea buckthorn (**Figure 4**). DSMs or DTMs are calculated by interpolation from points; they can be represented in vector mode as contour maps or in raster mode where each pixel contains elevation information. With a data point density of 1–2 points/m2 , there is no need to aim for a spatial resolution of DSM or DTM finer than 1 m/pixel. **Figure 4** shows the very wide foreshore along the port of Dunkirk, which gradually narrows toward the east. It also shows a morphology of bars and troughs on the foreshore. The sand and gravel spits are clearly visible.

#### *2.2.3 Airborne photogrammetry*

Photogrammetry is a nonactive telemetry process because no signal is emitted. It can be defined simply as the "science of making reliable measurements from photographs" [39]. It is most frequently applied to aerial photographs. It allows distance and surface measurements to be made in plan. It is also possible to extract altitude information from two vertical aerial photographs of the same area taken from different angles using the principle of stereoscopy.

Photogrammetry therefore requires a camera: formerly, a silver camera, and nowadays, a high-resolution digital sensor. These cameras are embarked on different types of carriers: airplanes [40], ultralight aircraft [41], or UAV [42–46]. The cameras are installed on stabilized gimbals that compensate for the vector's pitch and roll movements in order to minimize the blurring effects on the photographs.

"Conventional" airborne photogrammetry, using carriers flying at altitudes of 500–5000 m above the ground, provides results with highly variable vertical error margins that largely depend on the spatial resolution of the acquired images and therefore on the flying height. The margin of error can be metric (1.5 m in mountainous areas) [24] or decimetric (30 cm in coastal areas) [39], which is not accurate enough to detect fine detail and small amplitude changes on beaches.

In recent years, airborne photogrammetry by UAV at low altitude (below 150 m) and very high spatial resolution has rapidly gained popularity in coastal geomorphology. As mentioned above, this success has been made possible by the development of digital photogrammetry and the introduction of small, easy-touse, and affordable civilian UAVs. This success is also due to a technical change in photogrammetry, which was originally only an optical technique, requiring the use of bulky equipment and the permanent intervention of a technician, whereas it is now completely digital and can be automated almost all along the image processing workflow. Specialized software, but relatively easy to use for those who are familiar with geomatics, is available. These programs perform a correspondence analysis of the pixels of the images to "stitch" them together like a panorama; they calculate angles between the shooting points and these pixels in order to calculate depths of field or differences in altitude to represent the relief.

The protocol for organizing a UAV photogrammetry mission is simple. A flight plan is programmed, which forecasts the UAV's trajectory in the form of flight path parallel to each other in the area to be studied. The flight altitude of the aircraft is parameterized, which, depending on the sensor's fineness, determines the spatial resolution of the vertical images: the greater the height, the lower the resolution. For example, with a 20 MPixel sensor, a pixel of 2.41 μm, a focal length of 8.8 mm (DJI Phantom IV Pro or Mavic 2 Pro UAV), a flight height of 50 m, the spatial resolution of an image will be 1.4 cm/pixel. *P* × 10 \_

$$R = H \times 1000 \times \left[ \frac{P \times 10^{-4}}{F \times 10^{-1}} \right] \tag{1}$$

**99**

*Recent Advances in Coastal Survey Techniques: From GNSS to LiDAR and Digital…*

2.5 cm at altitude with a spatial resolution of 6 cm/pixel. The sharpness is such that when zooming in, wheel tracks of bulldozers replenishing the beach a little further east can be seen. Such spatial resolution of the DSM can be achieved, thanks to the very high density of the points. As with LiDAR data, photogrammetric measurements concern the surface of objects: the transformation of DSMs into DTMs is sometimes possible for spatially well-defined objects (buildings and small groves of vegetation) using the same methods. However, such high accuracy DTMs can hardly be obtained in densely vegetated areas. Interpretations of elevation changes must then be very cautious. An apparent increase in altitude may be the consequence of vegetation growth and not of sand accretion (see Section 3 of this chapter).

*Example of a very high resolution DSM computed from photogrammetric survey carried out with a DJI Mavic* 

A major interest of photogrammetry compared to LiDAR is that it allows the acquisition of aerial photographs on which many detailed observations can be distinguished (e.g., sand/vegetation limit on the upper beach and type of vegetation). These photographs, each covering a small area, can be assembled to form a full orthophotograph [39]. They are corrected for optical distortions using precisely known shooting parameters and recorded in the exif metadata of the files; distor-

GNSS, LiDAR, and photogrammetry can be classified in several ways. A distinction can be made between ground (GNSS and TLS) and airborne (LiDAR and photogrammetry) measurement methods, active (sending/receiving a signal; GNSS, TLS, and LiDAR) and nonactive (photogrammetry) measurement methods, in situ (GNSS) and telemetry (remote measurement, TLS, LiDAR, and photogram-

The quality of the results is closely linked to the positioning accuracy at the different stages of the technical protocols (e.g., TLS, vector, and ground control points) and therefore to GNSS, which remains an essential instrument for the

The graph in **Figure 6** aims at showing the complementarity of these techniques. Two parameters are taken into account: on the abscissa, the time scale/frequency of acquisition, and on the ordinate, the spatial density of the measurements, which is a key factor for the accuracy of DTMs or DSMs [47]. As the frequency of data acquisition increases and the degree of accuracy becomes more refined, finer analyses become possible. With regularly repeated high spatial resolution DSMs, it is thus

*DOI: http://dx.doi.org/10.5772/intechopen.91964*

*2 Pro UAV on the harbor of Dunkirk coast, northern France.*

tions due to relief are corrected using DSM.

metry) measurement methods.

calibration of the elevation data.

**2.3 Synthesis**

**Figure 5.**

where *R* is the spatial resolution of the image in cm/pixel, *H* is the flying height in meters, *P* is the sensor pixel size in microns, and *F* is the sensor focal length in millimeters.

Overlap between two successive photographs and lateral overlap between two parallel strips of photographs are programmed (80 and 60% are recommended, respectively). On the UAV's speed depends on these parameters. The speed must be slow enough to avoid blurring on the photographs. In the example shown in **Figure 5**, the flight height of the DJI Mavic 2 Pro was 65 m, and the speed was 5 m/s. Oblique aerial photographs can also be integrated into the data to be processed: they help to better model steep slope reliefs such as cliffs or foredune fronts. The flight plan is recorded and used for repeated overflights at different dates. During each campaign, ground control points, evenly distributed over the study area, are measured using GNSS. The coordinates of these control points allow the DSM to be set very precisely in plan and in altitude.

Digital photogrammetric processing with very high spatial resolution allows to generate a very high density of measurement points and to calculate very fine DSMs. The DTM, as shown in **Figure 5**, was calculated with an average density of 269 points/m2 . The average error margin of the measurements is 3.2 cm in the plane and

*Recent Advances in Coastal Survey Techniques: From GNSS to LiDAR and Digital… DOI: http://dx.doi.org/10.5772/intechopen.91964*

**Figure 5.**

*Spatial Variability in Environmental Science - Patterns, Processes, and Analyses*

enough to detect fine detail and small amplitude changes on beaches.

field or differences in altitude to represent the relief.

resolution of an image will be 1.4 cm/pixel.

set very precisely in plan and in altitude.

*R* = *H* × 100 ×

In recent years, airborne photogrammetry by UAV at low altitude (below 150 m) and very high spatial resolution has rapidly gained popularity in coastal geomorphology. As mentioned above, this success has been made possible by the development of digital photogrammetry and the introduction of small, easy-touse, and affordable civilian UAVs. This success is also due to a technical change in photogrammetry, which was originally only an optical technique, requiring the use of bulky equipment and the permanent intervention of a technician, whereas it is now completely digital and can be automated almost all along the image processing workflow. Specialized software, but relatively easy to use for those who are familiar with geomatics, is available. These programs perform a correspondence analysis of the pixels of the images to "stitch" them together like a panorama; they calculate angles between the shooting points and these pixels in order to calculate depths of

The protocol for organizing a UAV photogrammetry mission is simple. A flight plan is programmed, which forecasts the UAV's trajectory in the form of flight path parallel to each other in the area to be studied. The flight altitude of the aircraft is parameterized, which, depending on the sensor's fineness, determines the spatial resolution of the vertical images: the greater the height, the lower the resolution. For example, with a 20 MPixel sensor, a pixel of 2.41 μm, a focal length of 8.8 mm (DJI Phantom IV Pro or Mavic 2 Pro UAV), a flight height of 50 m, the spatial

[

where *R* is the spatial resolution of the image in cm/pixel, *H* is the flying height in meters, *P* is the sensor pixel size in microns, and *F* is the sensor focal length in

Overlap between two successive photographs and lateral overlap between two parallel strips of photographs are programmed (80 and 60% are recommended, respectively). On the UAV's speed depends on these parameters. The speed must be slow enough to avoid blurring on the photographs. In the example shown in **Figure 5**, the flight height of the DJI Mavic 2 Pro was 65 m, and the speed was 5 m/s. Oblique aerial photographs can also be integrated into the data to be processed: they help to better model steep slope reliefs such as cliffs or foredune fronts. The flight plan is recorded and used for repeated overflights at different dates. During each campaign, ground control points, evenly distributed over the study area, are measured using GNSS. The coordinates of these control points allow the DSM to be

Digital photogrammetric processing with very high spatial resolution allows to generate a very high density of measurement points and to calculate very fine DSMs. The DTM, as shown in **Figure 5**, was calculated with an average density of 269

. The average error margin of the measurements is 3.2 cm in the plane and

*P* × 10 \_ −4 *<sup>F</sup>* × 10−1

] (1)

Photogrammetry therefore requires a camera: formerly, a silver camera, and nowadays, a high-resolution digital sensor. These cameras are embarked on different types of carriers: airplanes [40], ultralight aircraft [41], or UAV [42–46]. The cameras are installed on stabilized gimbals that compensate for the vector's pitch and roll movements in order to minimize the blurring effects on the photographs. "Conventional" airborne photogrammetry, using carriers flying at altitudes of 500–5000 m above the ground, provides results with highly variable vertical error margins that largely depend on the spatial resolution of the acquired images and therefore on the flying height. The margin of error can be metric (1.5 m in mountainous areas) [24] or decimetric (30 cm in coastal areas) [39], which is not accurate

**98**

points/m2

millimeters.

*Example of a very high resolution DSM computed from photogrammetric survey carried out with a DJI Mavic 2 Pro UAV on the harbor of Dunkirk coast, northern France.*

2.5 cm at altitude with a spatial resolution of 6 cm/pixel. The sharpness is such that when zooming in, wheel tracks of bulldozers replenishing the beach a little further east can be seen. Such spatial resolution of the DSM can be achieved, thanks to the very high density of the points. As with LiDAR data, photogrammetric measurements concern the surface of objects: the transformation of DSMs into DTMs is sometimes possible for spatially well-defined objects (buildings and small groves of vegetation) using the same methods. However, such high accuracy DTMs can hardly be obtained in densely vegetated areas. Interpretations of elevation changes must then be very cautious. An apparent increase in altitude may be the consequence of vegetation growth and not of sand accretion (see Section 3 of this chapter).

A major interest of photogrammetry compared to LiDAR is that it allows the acquisition of aerial photographs on which many detailed observations can be distinguished (e.g., sand/vegetation limit on the upper beach and type of vegetation). These photographs, each covering a small area, can be assembled to form a full orthophotograph [39]. They are corrected for optical distortions using precisely known shooting parameters and recorded in the exif metadata of the files; distortions due to relief are corrected using DSM.

### **2.3 Synthesis**

GNSS, LiDAR, and photogrammetry can be classified in several ways. A distinction can be made between ground (GNSS and TLS) and airborne (LiDAR and photogrammetry) measurement methods, active (sending/receiving a signal; GNSS, TLS, and LiDAR) and nonactive (photogrammetry) measurement methods, in situ (GNSS) and telemetry (remote measurement, TLS, LiDAR, and photogrammetry) measurement methods.

The quality of the results is closely linked to the positioning accuracy at the different stages of the technical protocols (e.g., TLS, vector, and ground control points) and therefore to GNSS, which remains an essential instrument for the calibration of the elevation data.

The graph in **Figure 6** aims at showing the complementarity of these techniques. Two parameters are taken into account: on the abscissa, the time scale/frequency of acquisition, and on the ordinate, the spatial density of the measurements, which is a key factor for the accuracy of DTMs or DSMs [47]. As the frequency of data acquisition increases and the degree of accuracy becomes more refined, finer analyses become possible. With regularly repeated high spatial resolution DSMs, it is thus

**Figure 6.** *Spatial and temporal scales of application of the three techniques.*

possible to highlight otherwise undetectable detail changes at lower acquisition frequency and/or with coarser spatial resolution surveys (e.g., aeolian sand accumulations, see Section 2 of this chapter).
