1 of 18

Measuring wave runup and intertidal beach topography from online-streaming surfcam

Umberto Andriolo 1, Rui Taborda

Instituto Dom Luiz, University of Lisbon, Lisbon, Portugal

Elena Sánchez-García

Universitat Politècnica de València, Valencia, Spain

Abstract

In the swash zone, the variation of beach profile is fundamental to understand shore morphodynamics, while wave runup is a key parameter to evaluate coastal vulnerability to extreme events.

In such context, this work aimed to develop and validate two complementary methods to video-derive wave runup measurements and intertidal beach topography. Video data were obtained from a freely-available online-streaming surfcam installed at Costa da Caparica, monitoring Praia do Paraíso. RTK-GPS survey was performed to characterize the study area and to validate video-derived results.

Wave runup statistics (minimum, mean and maximum Rup) were obtained by statistical analysis of 144 Timestacks, with a total Root Mean Square Error (RMSE) of 0.18 m. Best detection performance was achieved for the Rupmean with an RMSE of 0.13 m.

Intertidal beach topography was carried out through estimating shoreline elevation by video observation. A vertical error of 0.18 m was obtained in computing 4 cross-shore beach profiles through Timestacks Applying the method to Variance, a video-derived Digital Terrain Model was implemented for an area of 11700 m2 with an elevation RMSE of 0.14 m.

This study successfully met the objective of developing simple techniques for swash zone hydro- and morphodynamic characterization. The outcomes provide new methods to study the dynamics induced by climate changes on littoral zone, including the impact of extreme events on coastal areas.

______________________ 1 Corresponding author: Umberto Andriolo, IDL, Universidade de Lisboa, Campo Grande, Building C6, 1749-016 Lisboa, Portugal. E-mail: uandriolo@fc.ul.pt

2 of 18

Introduction

The swash zone is the coastal region where the sea interacts with the land, considered between the run-down position to the limit of runup. In this area, the wave runup and the variation of beach profile are key parameters for evaluating coastal vulnerability to storm wave impact. Wave runup predictions are generally based on empirical formulations, whose performance depend upon limited wave and beach topographic data availability. Besides, standard traditional beach survey methods lack in space-time resolution, and require intensive human efforts.

The video monitoring technique has being proved as a valid methodology for analysing hydro- and morphodynamics of the swash zone. Argus monitoring program (Holman & Stanley 2007, for a comprehensive resume) has been providing coastal image data worldwide for the last two decades. In the early 2000’s, the improvement of the use of relatively cheap Internet Protocol cameras encompassed the expensive Argus installation and promoted the development of stand-alone systems (e.g., Cam-Era https://www.niwa.co.nz/our-services/online-services/cam-era; COSMOS, Taborda & Silva, 2012; CoastalComs, http://www.coastalcoms.com; Orasis http://www.vousdoukas.com/index_video.html; SIRENA, Nieto et al. 2010). The remote video imagery technique provides cost-efficient and high resolution data in time and space.

The coastal video-monitoring uses special optical products. Given a 10 minutes image sequence, Time exposure images (Timex) are composed by the averaged pixel intensity, while Variance images combine the standard deviation of pixel brightness. Timestack images are instead produced through sampling the pixel array corresponding to a cross-shore transect.

Despite the large exploitation of video imagery techniques, the use of online-streaming web-cam has been poorly investigated. For example, coastal “surfcams” are Internet Protocol cameras installed at the coast with the main aim of remotely providing visual information of sea state to surf users. To author’s knowledge, Bracs et al. (2016) and Mole et al. (2013) are the only available works in literature that investigated the use of such devices for qualitative nearshore studies. It is also important to note that such works used commercial software. Among the big amount of surfcams streaming coastal images worldwide, to date around 70 surfcams have being deployed by three different private companies on the Portuguese coast.

Background

Wave runup is defined as the upper limit of wave on the beach face. Wave runup determines sediment transport and defines storm impact on the coast. Empirical formulations of wave runup are functions of offshore wave parameters and beach slope. The leading edge of the swash is visible as wet–dry boundary on Timestacks. Several procedures have been proposed for automatizing the swash motion detection on Timestacks (e.g., Vousdoukas et al., 2012; Holland & Holman, 1993).

3 of 18

These approaches require sophisticated image processing and important computational time. In addition, automated processes are not always robust due to the fact that swash contour detection on the beach slope might be affected by poor image resolution and image noise caused by human beach use. Therefore, manual digitalization was often necessary to mark or to refine discrete runup positions (Huisman et al. 2011; Senechal et al., 2011; Ruggiero et al., 2004).

Recently, Simarro et al. (2015) proposed a method for deriving shoreline position and runup statistics based on the cross-shore variance function of a Timestack. The method relies on the identification of the minimum and the maximum pixel intensity variability for identifying the swash zone on the image. Shoreline position and R2% were derived by two empirical expressions, constrained by the use of two general coefficients derived by a small number of observations. Coefficients were varying for different beach slopes, tide elevation and breaking wave pattern characteristics. Therefore, there is still a need for a robust and simple universal technique for measuring wave runup through video data.

Shoreline position is defined by the boundary between water and dry sand (Boak & Turner, 2005). Previous works based the shoreline detection on wet-dry boundary on Timex (Osorio et al., 2012; Plant et al., 2007; Aarninkhof et al. 2003; Plant & Holman, 1997) and on Variance (Harley et al., 2013).

Shoreline elevation on beach slope was considered as the sum of tidal level, the wave-induced setup and swash-motion- induced height (Sobral et al., 2013; Vousdoukas et al., 2011; Plant et al., 2007; Aarninkhof et al., 2003). Wave runup modeling was applied to estimate such heights. Elevating shoreline position to such height over half of tidal cycle provided foreshore topography. Depending the classical shoreline elevation models on wave measurements and empirical formulations, there is a need for a simple shoreline height estimation based on observed data.

Given the abovementioned background, the objectives of this contribution are:

a) using online-streamed surfcam images for producing rectified Timex,

Variance and Timestack images;

b) proposing a simple methodology to measure wave runup statistics and

compute shoreline elevation through video observations;

c) applying the developed technique to Variance for characterizing wave

runup and beach intertidal topography of the video-monitored area.

4 of 18

Study Site

Costa of Caparica is a coastal stretch located on southern bank of the Tagus river inlet. The nearshore wave climate is characterized by significant wave heights Hs ranging from 0.5 m to 2.5 m, wave period T from 5 to 15 seconds, with higher frequencies and intensities coming from WSW and WNW. The tidal regime is semidiurnal with a maximum spring tide up to 3.8 m (Veloso-Gomes et al., 2004).

The study site is the Praia of Paraíso (38°38'30.5"N 9°14'21.1"W), one of the constrained beaches in front of Caparica village (Figure 1, a). The beach is around 1 km long, N-S oriented and bounded by rocky wall behind and two cross-shore artificial rocky groins, which extend seaward for 225 m (the northern) and 100 m (the southern).

Figure 1. a) Map of the study site location. The black box identifies the monitored area. The red dot indicates the camera position (Base map source: ESRI imagery). b) surfcam webpage from which the video data were retrieved.

Video data

The Surfline network (http://www.surfline.com) includes 10 video cameras, which are

installed with the main aim of remotely providing visual information of sea state to users along the Portuguese coast.

The surfcam mounted on the 8th floor of a hotel at Costa da Caparica was chosen for this work. Video data are streamed online at 25 frames per seconds (fps), and available online (Figure 1, b). The access to the installation site and the technical properties of the camera were denied by the company. Thus, camera position in real-world coordinates, camera internal and external parameters were unknown. Usually, the camera is set up for mechanically rotating and zooming to show different areas of the coast.

For this experimental work, we considered the data acquired by the video camera between 08:30 and 14:30 on 11th of November 2015, when the camera was set steady looking at the south part of Praia do Paraíso beach. The video data interval corresponded to increasing tide.

5 of 18

Methods

This section reports the proposed methods to: 1) compute precise repositioning of the camera, and surfcam images rectification; 2) detect wave swash and measure wave runup from Timestacks and 3) derive beach intertidal topography through Timestacks and Variance images. Figure 7 shows the flow chart of procedures followed for the aims.

Wave, tidal and topographic data

Wave data were retrieved from a wave model developed by the Portuguese Laboratory of Civil Engineering (LNEC) (http://ariel.lnec.pt). During the experiment, offshore significant wave height Hs was about 1.25 m, mean wave period Tm varied between 8 and 9 s, peak period Tp between 11 and 12.6 s.

Tidal data were obtained by the Cascais tide gauge (ftp://ftp.igeo.pt) deployed by the Portuguese Geographic Institute (IGEO). Tidal range was 2.3 m, with a minimum elevation related to the Mean Sea Level of -0.9 m at the start and 1.43 m at the end of the experience.

An RTK-GPS survey was performed to characterize the topography of the area on 12th of November 2015. Four cross-shore transects were surveyed during low tide in order to describe the beach slope from about -1 m to about 3.5 m, relatively to the Mean Sea Level. 39 Ground Control Points (GCPs) were also collected both on the dry beach and on structures to cover the image (Figure 2).

The foreshore slope ( ) was estimated with the best linear fit of the measured beach profile elevation between -0.5 m and 2 m. The beach slope for the 4 profiles was found to vary between 0.033 for the profile 1 and 0.031 for the profile 4, denoting morphological intermediate conditions. Considering wave characteristics and the mild beach slope (<0.1) during the experience, Praia do Paraíso was subjected plunging breakers (Iribarren & Nogales, 1949; Battjes, 1974)

Figure 2. Original frame extracted by video data. Blue dots indicate the 39 Ground Control Points collected by RTK-GPS for the rectification process. Red lines indicate the surveyed beach profiles, numbered from 1 to 4.

P11

P21

P31

P41

6 of 18

Video imagery rectification

The surfcam retrieved dataset consisted in 36 10-minutes videos, successively converted in a sequence of 21600 images (800 x 450 pixels resolution). Frames were extracted at 1Hz, a suitable frequency for the aim of this study.

The image dataset was rectified using the semi-automatic photogrammetric tool C-Pro (Sánchez-García et al., 2017, Sánchez-García et al., 2015). Selecting a minimum of 6 GCPs, C-Pro algorithms achieved the preliminary camera calibration and repositioning by Direct Linear Transformation relation. Successively, the external and internal photogrammetric orientation parameters were refined by an iterative collinearity adjustment to transform image data into real-world coordinates. In order to optimize the procedure, C-Pro used the terrestrial horizon line in the image as a geometric computational constraint. Finally, image rectification was performed by inverse mapping and nearest neighbor interpolation methods.

The rectified Timex and Variance images were produced over 10 minutes image sequence for the 6 monitored hours. The images were projected on the referenced plane identified by the 10-minutes-averaged sea level over the actual image sequence. Timestacks were composed by the time series of pixel intensity sampled along four cross-shore transects, corresponding to the four cross-shore profiles surveyed in the field (Figure 2). On such images, x-axes and y-axes represented cross-shore distance in meters and time in minutes, respectively.

a)

b)

c)

d)

Figure 3. Video imagery products. a) snapshot image; b) time-exposure (Timex) image; c) Variance image; d) Timestack image, produced sampling pixel array over an image sequence indicated by red line in a).

7 of 18

Wave runup

Simarro et al. (2015) pointed out that swash zone can be identified on Timestack through computing standard deviation along time-axis. The peak value coincides with the breaking wave position and the minimum value limit the dry beach. In fact, in the field of digital image processing, standard deviation is generally used to describe variation between neighbouring pixels (Madisetti et al., 1999). Thus, swash zone is identified between highest brightness (breaking waves) and lowest value (dry beach).

Hereafter, we refer to “wave swash” for the cross-shore location found on Timestacks (in pixel), to “wave runup” for the swash elevation after the interpolation with topographic survey (in meters). In addition, considering 10-minutes interval, we define as “minimum swash” the minima seaward position reached by the waves, while as “minimum runup” the minimum height on beach profile. Likewise, “maximum swash” and “maximum runup” outline maximum shoreward position and maximum height. Finally, mean swash and mean runup identify the averaged position and averaged height computed among all wave swash and runup occurred.

A simple method based on image statistical analysis is proposed for automatically deriving minimum, mean and maximum wave swash positions on Timestack. For the purpose, we introduced the dimensionless Coefficient of Variation (CV), defined as the ratio of standard deviation to the mean (Madisetti et al., 1999):

Eq. 1

where is the standard deviation and is its mean value.

The TimeStacK Method (TSKM) was built as follow: the standard deviation function was computed along time axis of the

Timestack, and normalized;

the coefficient of variation (CV) was computed;

the maximum swash position Swmax was found as the index of the

minimum value on the detrended CVd;

the minimum swash position Swmin corresponded to the index of the

maximum value on the detrended CVd;

the mean swash position Swmean was given by the index of the minimum

value among the absolute values of CVd, named CVdAbs (Figure 4).

Besides, each position of discrete wave swash Sw was manually digitalized on Timestacks following the standard procedure reported in literature (i.e., Vousdoukas et al., 2012). Runup measurements Rup were obtained by interpolating swash pixel coordinates with the topographic data, both for manual and video techniques (Figure 4).

8 of 18

Figure 4. Example of TSKM procedure. a) Timestack; white rectangle indicates the zone between Swmin and

Swmax; b) detrended CV (CVd) and its absolute values (CoVdAbs) of time-axis Timestack in a). Arrows and

text boxes indicate the derived swash positions Sw. Grey rectangle bounds the area between Swmin and

Swmax. The peak around x = 145 on CVd plot depends on noise tipically generated by human beach

occupation; c) beach surveyed profile corresponding to the transect covered by Timestacks. Arrows indicate

the process for deriving Rup through Sw found by CoV analysis.

In order to evaluate accuracy and goodness of the proposed technique, errors were defined as follow:

Eq. 2

Eq. 3

where the subscripts M and V stay for “manual” and “video-derived”, respectively. Root Mean Square Error (RMSE) was calculated for all the swash Sw (in pixel) and runup Rup (in meters) assessments (minimum, mean maximum positions and values):

√

∑

Eq.4

√

∑

Eq.5

where N is the number of Sw and Rup values considered.

9 of 18

Intertidal topography

Intertidal topography of the active beach profile was estimated through sequential combination of shoreline positions and water level over half of tidal cycle.

We propose a simple method based on the assumptions that minimum runup Rupmin identifies the tide water level ztide, and shoreline position (xSL) is equivalent to the mean swash position Swmean (Simarro et al., 2015; Huisman et al., 2011). Hereinafter regarding intertidal topography, we refer to “shoreline position” (xSL=Swmean) and to “shoreline elevation” (zSL).

The Shoreline Elevation Method (SEM) was based on wave runup observed on images. A first estimation of beach-face slope ( ) was carried out through a linear fitting of minimum swash positions Swmin elevated to tidal level ztide

(Figure 5,a). Given the cross-shore shoreline position xSL, the shoreline elevation zSEM relative to ztide was assessed multiplying the tangent of the fitting line by the cross-shore distance between shoreline positions xsl and minimum swash Swmin (Figure 5,b).

a)

b)

Figure 5. Procedure to derive zSEM. a) fitting of Rupmin to estimate β*. b) computation of zSEM through relation of Swmin, zSL and the tangent of β.

Shoreline elevation zSL was finally found as:

Eq. 6

The intertidal topography from Timestacks was computed using manually identified shoreline position xSL and computed zSL.

Accuracy estimation of the proposed technique was performed comparing shoreline elevations zSL with manual Rupmean M.

Eq.7

√

∑

Eq. 8

Beach-face slope ( ) obtained by the intertidal topography was verified against RTK-GPS surveys. Slopes were assessed through the best linear fitting of shoreline elevations comprised within three different elevation intervals, namely [-0.5 , 2] m , [0 , 2] m and [0.5 , 2] m.

10 of 18

a)

b)

Figure 6. a) original frame of 11:00:01. Coloured lines represent surveyed profiles. Black lines indicate the

10 profiles sampled on Variance. Dashed cyan lines delimitate the rectified area shown in b). Axis units are

in pixels. b) rectified Timex produced by image sequence 11:00 ÷11:10. See legend in a) for lines

explanation. Coordinates in ETRS 89 – PT-TM06 system. Timex is used just for better representation.

The standard deviation profile of Timestack coincides with the same profile sampled on Variance (Simarro et al., 2015). Therefore, the same method used for deriving wave swash statistics on Timestack was applied to 10 parallel cross-shore profiles sampled on the 36 rectified Variance (Figure 6). For each profile, shoreline positions xSL and elevation zSL were automatically computed to carry out a Digital Elevation Model (DEM) of the monitored area. Video-derived DEM was compared to beach surface obtained by the topographic survey.

Figure 7: Webcam images analysis flow chart of TimeStacK Method (TMSK) and Shoreline Elevation

Method (SEM).

11 of 18

Results

Rectification

The overall positional and rectification accuracy obtained by C-Pro was evaluated comparing the real-world coordinates of 39 GCPs against their projected coordinates.

Figure 8 shows the precision of C-Pro rectification process. Most of the checkpoints were reprojected within 1 m, suitable value for this study. Maximal errors (~ 2.5 m) were found for the most distant points to the camera.

a)

b)

Figure 8. a) Portion of original snapshot. Blue dots indicate surveyed GCPs white dots represent their

projected position. b) Rectified image where the coloured points indicate E-N positioning error in m for

each of the GCPs. Coordinates in ETRS 89 – PT-TM06 system.

Wave runup

A total of 144 Timestacks were processed to automatically detect swash positions Sw and runup Rup.

Figure 9. Example of wave swash detection results and comparison against manually digitized positions. See

Figure 4 for Timestack details.

-9.65 -9.645 -9.64 -9.635 -9.63 -9.625 -9.62

x 104

-1.1355

-1.135

-1.1345

-1.134

-1.1335

x 105

Easting coordinates

No

rth

ing

co

ord

ina

tes

<0.5 m

0.5 1 m

1 1.5 m

> 1.5 m

12 of 18

Table 1 reports the comparison between manual and automated detection for minimum, mean and maximum values, respectively. Total values were obtained averaging the errors resulted for the singles profiles.

Table 1. RMSE error obtained by wave swash and wave runup detection algorithm. Total RMSE report the average error for the 4 profiles

Swmin

(pixels)

Swmean

(pixels)

Swmax

(pixels)

Rupmin

(m)

Rupmean

(m)

Rupmax

(m)

Total RMSE

11.6 7.15 7.28 0.19 0.136 0.229

Profile 1 8.2 7.0 4.5 0.203 0.166 0.159

Profile 2 8.4 5.9 3.7 0.171 0.203 0.184

Profile 3 15.2 4.9 7.5 0.203 0.125 0.240

Profile 4 12.7 9.8 11.3 0.230 0.161 0.349

Regarding automated detection of Swmax, a RMSE of about 7 pixels led to a RMSE of about 0.3 m for video-derived Rupmax. The image processing algorithm occasionally returned rough measurements when human occupation on the beach affected the automated method. Worst accuracy was registered on steeper beach slope, where small horizontal errors determined greater elevation imprecision.

Swmin estimations returned an RMSE of 11.4 pixels, while RMSE for Rupmin was 0.19 m. During low tide, Swmin identification was difficult due to the dark colour of saturated beach, as in Simarro et al. (2015). Moreover, transect length chosen for Timestack production was too short to entirely show the swash zone during low tide. Therefore, Swmin detection was affected by the lack of wave development. On the contrary, Swmin was well detected for higher water levels and steeper beach slope.

Automated determination of Swmean and Rupmean showed the best accuracy, with a RMSE of 6.7 pixels and 0.13 m, respectively. As the proposed method bounds the search for Swmean between related Swmin and Swmax, imprecision was mainly generated by the errors that were previously discussed. Figure 9 shows single example of detected swash positions, in which Swmean is well detected despite low accuracy in identifying Swmin and Swmax.

Overall, the comparison between manual and automated technique for 432 Rup values (minimum, mean and maximum) resulted in an average RMSE of 0.184 m.

13 of 18

Intertidal topography

Figure 10 shows the intertidal beach topography assessed through Timestack analysis. In order to evaluate the Shoreline Elevation Method (SEM) accuracy, the shoreline positions xSL were elevated to zSL and plotted against the surveyed beach profile. The RMSE varied from a minimum of 0.139 m for profile 2 to a maximum of 0.193 m for the profile 3.

In general, intertidal topography carried out by SEM performed well for ztide higher than the Mean Sea Level (z=0), while it was overestimated for lower tide level. Nevertheless, errors were comparable with more sophisticated state-of-art shoreline elevation models (Sobral et al., 2013; Vousdoukas et al., 2011; Plant et al., 2007; Aarninkhof et al., 2003).

a)

b)

c)

d)

Figure 10. a), b) c) d) report the produced intertidal topography for the profiles 1,2,3,4, respectively. Blue squares indicate shoreline derived by Shoreline Elevation Method (SEM). For comparison, manual Rupmean is shown in red diamonds.

Results of SEM for zSL were also compared to the manual-derived Rupmean (Figure 11). Considering all the 144 measures used to assess the intertidal topography, the total RMSE was 0.18 m. Maximum difference between Rupmean and zSL results was 0.45 m, median value among all measures was 0.028 m. The accuracy was poorer during low tide, when SEM was sensitive to dissipative conditions and saturated beach.

14 of 18

Figure 11. Scatter plot of zSEM, computed by the Shoreline Elevation Method (SEM), against shoreline manual elevation (Rupmean).

Figure 12 shows the comparison between surveyed and SEM-derived beach-face slopes β for three different elevation intervals. On profile 1, the slope was overestimated for the all three slope intervals considered, with a maximum error of 0.014 and a mean error of 0.09. For the profiles 3 and 4, the SEM tended to underestimate the beach-face gradient, however the minimum accuracy was of 75%. Finally, best estimation by SEM was found for the profile 2, with a maximum error of 0.003. Error magnitudes were comparable with the ones obtained by the fully automated technique proposed by Vousdoukas et al. (2011).

Figure 12. Comparison between surveyed profile (filled circles) and video-derived (squares) beach-face

slope, expressed as tangent of the angle formed with the horizontal plane. Colours indicate profiles.. On Y-

axis, values are referred to the tangent of the slope. On X-axis, the beach elevation intervals considered for

the analysis.

15 of 18

Variance images

The video-derived DEM is shown against the surveyed surface in Figure 13. The beach surface was computed for an area of around 130 m cross-shore and 90 m along-shore, with an elevation range that went from -1.2 m to 1.5 m. Over a total area of about 11700 m2, average RMSE in elevation was 0.14 m, with a maximum of 0.28 m (1 m2 resolution grid).

a)

b)

c)

d)

e)

f)

Figure 13. Intertidal Digital Elevation Model. On left column (a, c, e), beach surface derived by RTK-GPS

survey. On right column (b, d, f), video-derived DEM from Variance. On first row (a, b) DEMs are plotted

on original frames; on second row (c, d) on rectified images, on third row (e, f) the 3D plots. Colorbars for

elevation are common to rows.

The main error was introduced by the inaccuracy in minimum wave swash detection during low tide. Nevertheless, different errors might also be induced by tidal measurements, image resolution, repositioning error during rectification.

16 of 18

Conclusions

This paper has reported efforts to take advantages of online video streaming surfcam for assessing wave runup measurements and intertidal beach topography.

The adoption of surfcam network existing and wordwile present coastal video infrastructure is an attractive solution for supporting coastal change monitoring and coastal management. The C-Pro tool provide rectified surfcam images, retrieved online with accuracy comparable to the classic phogrammetric techniques, allowing quantitative hydro- and mophodynamic analysis of the monitored area.

Wave runup statistics was successfully achieved by a soft-computing algorithm based on statistical properties of Timestack. The technique is easy to replicate and its implementation can improve automated runup measurements and parameterization. It is important to note that the method can be also applied to Variance images for studying wave runup along-shore variability.

The implementation of a new method for shoreline elevation returned an adequate representation of beach slope and intertidal beach topography. Although the technique needs to be applied to sites with different characteristics for a universal validation, data results showed similar accuracy to more sophisticated models and much potential.

We highlight the fact that the methods proposed for easyly runup measurements and beach intertidal topography estimation can be applied on Variance images, which have been storing by dozens of coastal video systems worldwide for the last two decades.

Acknowledgment

Umberto Andriolo was supported by the EARTHSYSTEM Doctorate Programme led by Institute Dom Luiz Associate Laboratory at the University of Lisbon (SFRH/BD/52558/2014). Elena Sánchez-García is supported by the Spanish Ministry of Education, Culture and Sport grant (state program in I+D+i 2013–2016).

17 of 18

References

Aarninkhof, S. G., Turner, I. L., Dronkers, T. D., Caljouw, M., & Nipius, L., 2003. A video-based technique for mapping intertidal beach bathymetry. Coastal Engineering, 49(4), 275-289. doi:10.1016/s0378-3839(03)00064-4

Battjes, J., 1974. Surf Similarity. Coastal Engineering 1974. doi:10.1061/9780872621138.029

Boak, E. H., Turner, I. L., 2005. Shoreline Definition and Detection: A Review. Journal of Coastal Research, 214, 688-703. doi:10.2112/03-0071.1

Bracs, M. A., Turner, I. L., Splinter, K. D., Short, A. D., Lane, C., Davidson, M. A., … Cameron, D., 2016. Evaluation of Opportunistic Shoreline Monitoring Capability Utilizing Existing “Surfcam” Infrastructure. Journal of Coastal Research, 319, 542-554. doi:10.2112/jcoastres-d-14-00090.1

Harley, M. D., Andriolo, U., Armaroli, C., & Ciavola, P., 2013. Shoreline rotation and response to nourishment of a gravel embayed beach using a low-cost video monitoring technique: San Michele-Sassi Neri, Central Italy. J Coast Conserv, 18(5), 551-565. doi:10.1007/s11852-013-0292-x

Holland, K. T., Holman, R. A., 1993. The statistical distribution of swash maxima on natural beaches. Journal Geophysical Research, 98(C6), 10271. doi:10.1029/93jc00035

Holman, R., Stanley, J., 2007. The history and technical capabilities of Argus. Coastal Engineering, 54(6-7), 477-491. doi:10.1016/j.coastaleng.2007.01.003

Huisman, C. E., Bryan, K. R., Coco, G., & Ruessink, B., 2011. The use of video imagery to analyse groundwater and shoreline dynamics on a dissipative beach. Continental Shelf Research, 31(16), 1728-1738. doi:10.1016/j.csr.2011.07.013

Iribarren, C.R., Nogales, C., 1949. Protection des Ports. XVIIth International Navigation Congres, Section II, Communication, 31-80 Ibaceta, R., Almar, R., Lefebvre, J-P., S

Madisetti, V., Williams, D. B., Chapman and Hall, & CRC Press., 1999 . Digital signal processing handbook: CRCnetBASE 1999. Boca Raton, Fla.: Chapman & Hall/CRCnetBase.

Mole, M. A., Mortlock, T. R., Turner, I. L., Goodwin, I. D., Splinter, K. D., & Short, A. D., 2013. Capitalizing on the surfcam phenomenon: a pilot study in regional-scale shoreline and inshore wave monitoring utilizing existing camera infrastructure. Journal of Coastal Research, 165, 1433-1438. doi:10.2112/si65-242.1

Nieto, M. A., Garau, B., Balle, S., Simarro, G., Zarruk, G. A., Ortiz, A., … Orfila, A., 2010. An open source, low cost video-based coastal monitoring system. Earth Surf. Process. Landforms, 35(14), 1712-1719. doi:10.1002/esp.2025

Osorio, A., Medina, R., & Gonzalez, M., 2012. An algorithm for the measurement of shoreline and intertidal beach profiles using video imagery: PSDM. Computers & Geosciences,46, 196-207. doi:10.1016/j.cageo.2011.12.008

18 of 18

Plant, N. G., Aarninkhof, S. G., Turner, I. L., & Kingston, K. S., 2007. The Performance of Shoreline Detection Models Applied to Video Imagery. Journal of Coastal Research, 233, 658-670. doi:10.2112/1551-5036(2007)23,2.0.co;2

Plant, N. G., Holman, R. A., 1997. Intertidal beach profile estimation using video images. Marine Geology, 140(1-2), 1-24. doi:10.1016/s0025-3227(97)00019-4

Ruggiero, P., 2004. Wave run-up on a high-energy dissipative beach. J. Geophys. Res, 109(C6). doi:10.1029/2003jc002160

Sánchez-García, E., Pardo-Pascual, J.E., Balaguer-Beser, A., Almonacid-Caballer, J., 2015. Monitorización de espacios costeros mediante un sistema fotogramétrico: C-Pro. XVI Congreso de la Asociación Española de Teledetección “Teledetección: Humedales y espacios protegidos”; Sevilla, España. ISBN: 978-84-608-1726-0

Sánchez-García, E., Balaguer-Beser, A., Pardo-Pascual, J. E., 2017. CPro: A Coastal Projector monitoring system using terrestrial photogrammetry with a geometric horizon constraint. ISPRS Journal of Photogrammetry & Remote Sensing (submitted)

Senechal, N., Coco, G., Bryan, K. R., & Holman, R. A., 2011. Wave runup during extreme storm conditions. J. Geophys. Res, 116(C7). doi:10.1029/2010jc006819

Simarro, G., Bryan, K. R., Guedes, R. M., Sancho, A., Guillen, J., & Coco, G., 2015. On the use of variance images for runup and shoreline detection. Coastal Engineering, 99, 136-147. doi:10.1016/j.coastaleng.2015.03.002

Sobral, F., Pereira, P., Cavalcanti, P., Guedes, R., & Calliari, L., 2013. Intertidal Bathymetry Estimation Using Video Images on a Dissipative Beach. Journal of Coastal Research,165, 1439-1444. doi:10.2112/si65-243.1

Taborda, R. and Silva, A., 2012. COSMOS: A lightweight coastal video monitoring system. Computers & Geosciences, 49, 248–255. doi.org/10.1016/j.cageo.2012.07.013

Veloso-Gomes, F., Pinto, F. T. & Barbosa, J. P., 2004. Rehabilitation study of costal defense Works and artificial sand nourishment at Costa da Caparica, Portugal. Proceedings of the 29th International Conference of Coastal Engineering, pp. 3429-3440

Vousdoukas, M. I., Ferreira, P. M., Almeida, L. P., Dodet, G., Psaros, F., Andriolo, U., Ferreira, Ó. M., 2011. Performance of intertidal topography video monitoring of a meso-tidal reflective beach in South Portugal. Ocean Dynamics, 61(10), 1521-1540. doi:10.1007/s10236-011-0440-5

Vousdoukas, M. I., Wziatek, D., & Almeida, L. P., 2012. Coastal vulnerability assessment based on video wave run-up observations at a mesotidal, steep-sloped beach. Ocean Dynamics, 62(1), 123-137. doi:10.1007/s10236-011-0480-x