Low-budget topographic surveying comes of age: Structure from motion photogrammetry in geography and the geosciences

Abstract

I Introduction

Photogrammetry is a technique by which volumetric information describing surface structures can be obtained by exploiting parallax – that is, the difference in the apparent position of an object, given the varying perspective provided by overlapping images captured from different viewpoints. Structure from motion (SfM) photogrammetry uses advances in computer vision algorithms to implement and expedite the digital photogrammetric workflow. Consequently, SfM photogrammetry can work with un-registered image sets captured from un-calibrated consumer cameras, although users must be mindful of the impacts of camera settings on the quality of resultant data and products (O’Connor et al., 2017). Until quite recently, topographic point cloud data were generally collected and delivered to science users by agencies and organisations with access to high-budget equipment (e.g. laser scanners on board piloted aircraft), and the raw data required powerful computers for data processing. These logistical requirements put such data out of the reach of some users (Westoby et al., 2012). The emergence of SfM photogrammetry as a low-cost surveying tool has, in contrast, provided a more democratic means by which scientists can capture and process their own point cloud data, and, resultantly, SfM photogrammetry now represents a core data capture and analysis approach within the environmental and geosciences disciplines. The paradigm for geospatial surveying is now shifting so that scientists themselves are the data suppliers (Garrett and Anderson, 2018); armed with consumer-grade cameras, it is possible to gather data and then process those data using SfM approaches to answer a range of environmental science questions.

Applications for SfM photogrammetry to date have included characterisation of forests (Dandois and Ellis, 2010; Mlambo et al., 2017; Zahawi et al., 2015), rivers (Marteau et al., 2017; Woodget et al., 2015), snow (Nolan et al., 2015) and coral reef systems (Casella et al., 2017; Leon et al., 2015), to name just a few. SfM has also been applied to analyse archival image data successfully – for example, for evaluating the dynamism of river floodplains (Bakker and Lane, 2017). However, this new-found route for self-service geospatial and volumetric data is paved with complexities for the user. Data can be captured from ground-based perspectives or from aerial viewpoints (drone and kite platforms facilitate this greatly), and there are a multitude of free, open-source or commercial software products available with which to process the photographic data into point clouds. Evaluating the quality of the resulting orthomosaics, point clouds and products requires independent data from Global Navigation Satellite Systems (GNSS) or other systems, which may have a much higher unit cost than the equipment used to capture the original photographic data.

Implementing the SfM workflow can be computationally expensive for large projects comprising multiple thousands of images, requiring users to have access to high-performance computing, or at the very least a powerful desktop with a high-specification graphics processing unit and large hard disk. Techniques for a detailed uncertainty assessment of such data can be memory heavy, and have yet to be widely adopted (Dall’Asta et al., 2015; James et al., 2017b; Murtiyoso et al., 2018).

In this editorial, which introduces a special issue on the topic of SfM photogrammetry in geography and environmental science, we will provide a brief summary of the status quo of SfM photogrammetry science, through a lens on published work. We begin by evidencing the rise of SfM photogrammetry work in geography and the geosciences, after which we summarise the main technical advances that have led to this expansion and uptake of the method. We then discuss how SfM photogrammetry has benefited geomorphology and natural hazards research, alongside ecology and hydrology research, throwing a spotlight on the papers contained within this special issue as we proceed.

II The rise of SfM photogrammetry within geography and geosciences

Within geographical and environmental science disciplines, there has been a recent upsurge in scholarly work that has developed, applied, evaluated and validated SfM photogrammetry approaches. A SciVal (www.scival.com) literature search on 1 January 2019 (looking for the terms “structure from motion” or “structure-from-motion” in the title, abstract or keywords), which returned 605 journal articles and review papers published up to 2018 within the Environmental, Agricultural, Biological, Earth and Planetary sciences, demonstrates the rapid, recent growth in this field (Figure 1).

We draw the reader’s attention to several key review papers (Bemis et al., 2014; Eltner et al., 2016) as well as three “early adopter” papers (Fonstad et al., 2013; James and Robson, 2012; Westoby et al., 2012), each of which provide an excellent starting point from which to understand the technicalities and capabilities of SfM photogrammetry. Recent refinement of technical SfM workflows (Carbonneau and Dietrich, 2017; James et al., 2017a, 2017b; Turner et al., 2012) has been paralleled by considerable advances in applications, including in spatial ecological (Cunliffe et al., 2016) and geomorphological (Javernick et al., 2014) areas of research, extending to the evaluation of plot-to-landscape scale processes using SfM applied to data acquired from different platforms (Smith and Vericat, 2015). The coincident rise of lightweight drone technology (sometimes referred to as “unmanned aerial vehicles” (UAVs)) alongside that of SfM photogrammetry has fuelled further the upsurge in interest and use of the technique within natural sciences (Anderson and Gaston, 2013). From these papers, one can get a flavour for the breadth of work within this emerging discipline of self-service, fine spatial resolution surveying.

III Technical advances

The contribution of SfM photogrammetry to environmental and geographical research has been supported by advances across image-acquisition platforms, cameras and processing software. Research questions have been addressed using image data collected from the ground to space, but aerial views from both piloted platforms (e.g. gyrocopters, parascenders, helicopters and fixed-wing aircraft) or unpiloted systems (e.g. kites, rotary or fixed-wing drones) are being increasingly used. In particular, the rapid advances in lightweight drone technology have increased the availability of data with a combination of centimetric spatial resolution and up to kilometric spatial coverage.

Drones used for geographical research have evolved from generally bespoke or kit-built systems requiring manual control by an experienced pilot (e.g. Niethammer et al., 2010), to off-the-shelf aircraft that are capable of near-autonomous surveying (e.g. Nakano et al., 2014). This transition has enabled image collection to advance from the characteristically irregular acquisition of manually controlled flights, to being fully autopilot-controlled and Global Positioning System (GPS)-synchronised, following programmed survey designs and flight lines. Simultaneously, image acquisition systems have evolved from typically using consumer cameras individually mounted into airframes (e.g. compact cameras and digital Single Lens Reflex (SLR)), to integrated, gimbal-mounted lightweight cameras (e.g. DJI’s 1-inch 20-megapixel CMOS sensor: https://www.dji.com/phantom-4-pro) and specialised imaging systems including thermal (e.g. the FLIR Duo Pro thermal sensor for drones: https://www.flir.co.uk/products/duo-pro-r/; and the Workswell WIRIS mini: https://www.workswell-thermal-camera.com/wiris-mini/) and multispectral (e.g. the Parrot Sequoia multispectral sensor) sensors.

Parallel advances have been made in image processing. Initial studies were carried out using early, widely available SfM-based software that was either web-based (e.g. PhotoSynth; Dandois and Ellis, 2010; Dowling et al., 2009; Rosnell and Honkavaara, 2012; Stimpson et al., 2010) or run on local PCs (e.g. Bundler with patch-based multi-view stereo; Castillo et al., 2012; James et al., 2012; Niethammer et al., 2010; Turner et al., 2012; Welty et al., 2010). With a heritage in computer science, such software tended to emphasise processing speed over metric accuracy, with aspects such as georeferencing left to external processing. As the broad utility of SfM photogrammetry became apparent, software began to include many more aspects of rigorous photogrammetry, such as integrating georeferencing directly into the workflow (e.g. Agisoft PhotoScan: www.agisoft.com). Drone-focussed software is also now commonly used (e.g. Pix4D: https://www.pix4d.com; and DroneMapper: https://dronemapper.com/), and these can seamlessly integrate survey design, flight planning and image processing, including radiometric considerations for multispectral data.

Following these recent developments in SfM workflows, we see two assessments of techniques within this special issue. Firstly, Griffiths and Burlingham compare the results of processing a salt marsh drone survey with different software and different approaches to camera calibration. Through assessing the resulting systematic error within their model, they show that camera self-calibration can either under-perform or exceed the performance of a pre-calibrated camera model, depending on the software used. Their work underscores some of the complexities involved in SfM photogrammetry and the importance of carefully designed error checks. Also in this issue, Ratner et al. return to the driver behind the early advances in automating SfM by exploring the use of crowd-sourced imagery to generate three-dimensional (3D) surface models. Using a volcanic crater as a case study location, they show that data acquired by volunteers walking around the area can be used to generate topographic data suitable for use in disaster risk reduction scenarios.

IV Geomorphology, landscape evolution and natural hazards research

The field of geomorphology has benefited strongly from the emergence of the current generation of SfM tools and associated surveying platforms. The versatility of SfM and the capacity for specialists and non-specialists alike to generate repeat, high-resolution topographic datasets lends itself naturally to geomorphological investigation. Accordingly, uptake by geomorphologists and those working on landscape evolution problems more broadly has been particularly rapid and proactive.

Within the field of fluvial geomorphology, SfM has enabled advances in both fluvial landform and landscape mapping at the reach (Dietrich, 2017; Javernick et al., 2014; Woodget et al., 2015), landform (Vázquez-Tarrío et al., 2017) and micro-scales, such as within experimental flumes (Morgan et al., 2017). SfM datasets have also informed fluvial process analysis; for example, work by Prosdocimi et al. (2017) demonstrates the utility of 4D analysis of SfM datasets for quantifying spatiotemporal patterns of fluvial erosion and deposition using smartphone cameras. Even in the absence of repeat datasets, SfM topography can inform hydrological process analysis; Smith et al. (2014) extracted high water marks from SfM-derived topography and used this information in combination with 2D hydraulic modelling to reconstruct reach-scale peak flow magnitudes in a flash-flood-affected catchment.

In coastal and tidal environments, SfM has been applied to the reconstruction of beach (Brunier et al., 2016), dune (Duffy et al., 2018; Mancini et al., 2013), cliff (James and Robson, 2012; Ružić et al., 2014) and rocky shore platform (Cullen et al., 2018) geomorphic structures. Work on tidal wetlands has shown that the technique can reveal geomorphic features that would be otherwise unresolvable in aerial Light Detection and Ranging (LiDAR)-derived data (Kalacska et al., 2017). Cullen et al. (2018), for example, used SfM-derived models to retrieve the volumes of percussion marks caused by clast abrasion on rocky shore platforms, whilst also advocating for the use of SfM photogrammetry to bridge scale-dependent methodological and observational constraints – a theme which has emerged across disciplines.

The uptake of SfM methods by cryospheric researchers has led to a number of notable advances. Ryan et al. (2015) were among the first to apply SfM photogrammetry to image data acquired from a long-range drone to provide insights into the calving dynamics of a large tidewater glacier at fine spatiotemporal resolution, whilst other notable contributions – for instance, Mallalieu et al. (2017) – have pioneered the application of terrestrial SfM for reconstructing ice-margin and ice-cliff dynamics, providing insights into both the seasonality and spatial variability of glacier surface evolution at the meso-scale. Similarly, Immerzeel et al. (2014) have applied SfM techniques for monitoring dynamics of high-altitude, debris-covered glaciers in the Himalaya, based on drone-captured images. At the micro-scale, work by Kääb et al. (2014) has shed new light on the surface evolution of periglacial sorted circles over multi-annual timeframes.

Some recent studies have also explored the potential for applying SfM methods to archival (predominantly aerial) photosets, and employing 4D analysis to quantify historic landscape evolution, including mountain glacier extent (Midgley and Tonkin, 2017; Mölg and Bolch, 2017) and braided river-floodplain systems (Bakker and Lane, 2017). Quite correctly, many of these studies advocate for the careful consideration of systematic error propagation when applying SfM reconstruction to imagery that has not been acquired with this express purpose in mind.

In this issue, Mather et al. compare the utility of archive aerial LiDAR topography and aerial photography, and aerial photography and topographic models acquired through SfM processing of drone-captured photosets for the recognition and automated mapping of periglacial features on an upland site in south-west England. Significantly, the authors develop an integrated approach to landform identification which utilises coarse-spatial resolution image data for landform mapping (<100 m scale), augmented by the use of fine-spatial resolution data for identifying smaller landform elements such as boulders (<1 m). In its broader sense, the work advocates for careful consideration of the appropriate spatial scale of remote-sensing-derived image data and topography with respect to the scale of the landscape features under investigation. Also in this issue, Derrien et al. apply SfM photogrammetry to aerial photosets of Piton de la Fournaise, one of the world’s most active volcanoes, following a summit collapse in 2007. In an excellent example of the use of SfM to inform hazard exposure at a popular tourist geosite, the authors applied elevation model differencing and 2D feature-tracking, respectively, to SfM-derived models and orthophotographs of the caldera in 2008 and 2015 to develop a comprehensive picture of mass wasting processes and ground motion across the site. The authors identified retrogressive erosion of the caldera rim caused by the widening of ground fractures, including an increase in widening rates since volcano reactivation in 2014, as well as extensive rock slope deformation and debris avalanche activity. By classifying the wider caldera rim area according to the magnitude and type of ground deformation or mass wasting hazard, and human exposure, the work clearly demonstrates the value of using 4D analysis of repeat SfM topography to inform risk analysis.

V Ecological and hydrological research

Mirroring advances within geological and geomorphological disciplines, ecological and hydrological research has exhibited a similar increase in uptake in the use of SfM photogrammetry approaches. This is because spatial datasets that allow structure/function relationships to be explored and have the potential to deliver a step-change in scientific insight within these disciplines.

Within terrestrial vegetation ecology, the pioneering work by Dandois and Ellis (2010) was amongst the first to demonstrate the potential information content of SfM-derived point clouds, demonstrating how fine spatial resolution measurements of vegetation structure could be applied to the assessment of biomass, carbon and, in forestry, fire and land management applications. Since this early paper (which utilised its own open-source Ecosynth SfM-based software: http://ecosynth.org/), there has been a significant adoption of SfM approaches within spatial ecology. There are plentiful examples of SfM being used for tree-height inventory for forestry applications (e.g. Birdal et al., 2017) and for estimating stem parameters for timber valuation purposes (e.g. Mikita et al., 2016), as well as for biomass inventory in tropical forests (e.g. Messinger et al., 2016). There has also been considerable exploration of the value of SfM approaches in agricultural settings, for delineating individual trees (Balsi et al., 2018; Ok and Ozdarici-Ok, 2018), for crop height and growth rate estimation in wheat crops (Holman et al., 2016) and for measuring grassland sward-height spatial variability (Forsmoo et al., 2018). In shrub-dominated systems where vegetation exhibits a shorter sward, there are also benefits to the SfM approach; Olsoy et al. (2018) demonstrate how SfM point clouds deliver useful information on habitat heterogeneity, and they argue that fine-grained information can improve decision-making for informing conservation practices. Working in drylands, Cunliffe et al. (2016) have also demonstrated the application of SfM-derived point clouds for above-ground biomass estimation. Another recent study by Webster et al. (2018) has pioneered the use of coincident capture and SfM processing of optical and thermal imagery from a UAV to quantify the 4D evolution of forest canopy temperatures, with potential utility for spatial and volumetric microclimate evaluation. Critically, we note the work of Wallace et al. (2016), who show that in complex forest systems, airborne laser scanning (ALS) may deliver comparatively more accurate estimates of the vertical structure of forests compared to SfM products. However, both Wallace et al. (2016) and Mlambo et al. (2017) comment on the adequacy of SfM products as a lower-cost alternative to ALS for surveying forest stands – which is particularly relevant for those working in remote regions of the world where access to piloted aircraft with ALS equipment is not feasible due to accessibility or cost (Messinger et al., 2016).

Beyond terrestrial ecology, there are recent examples of SfM photogrammetry being implemented in coastal areas, for example, in assessing the climate change impacts of sea-level rise on turtle nesting habitat (e.g. using SfM to deliver an accurate fine-grained model of beach topography; Varela et al., 2019), and for evaluating coral reef structures from both an airborne perspective (e.g. Casella et al., 2017) and from close-range underwater photography (Burns et al., 2015; Leon et al., 2015; Storlazzi et al., 2016; Bayley et al. (2019)).

At the interface of ecology and hydrology, Mercer and Westbrook (2016) have demonstrated the utility of SfM photogrammetry for delivering spatial and volumetric information about complex microform topography, which is so important in defining the ecohydrological function of complex peatland and wetland systems. In tidal wetlands, a similar relationship exists between eco-morphological structure and hydrological function, and Kalacska et al. (2017) demonstrate the successful application of SfM photogrammetry to the characterisation of critical hydrological features, including creeks and pond connectivity.

In hydrology, there are also plentiful examples of SfM-derived data being used to deliver new understanding – we refer readers to some of the examples given in section 4. Prosdocimi et al. (2015) showed how SfM-derived elevation models generated using images from a consumer-grade smartphone could deliver a quantitative estimation of deposition and erosion volumes, with reasonable correspondence to those obtained from terrestrial laser scanning (TLS) methods. They extended this method later to measure hydrologically eroded volumes of soil within a vineyard system with good efficacy (Prosdocimi et al., 2017). Importantly, in the context of erosion studies, Smith and Vericat (2015) urge users to exercise caution when scaling up SfM experiments over larger extents – although some of the errors they highlight can now be mitigated through rigorous methodological steps or accounted for in post-processing using sophisticated error evaluation techniques that are now available (e.g. Monte Carlo point-based uncertainty estimation; James et al., 2017b). Castillo et al. (2012) and Glendell et al. (2017) both deliver important work that evidences the cost-effectiveness of SfM erosion estimates over other volumetric surveying techniques.

In this issue, Neverman et al. evaluate results obtained from TLS and SfM photogrammetry surveys captured over an exposed gravel bar in the gravel-bedded Pohangina River, at three points in time, between which the bar had been inundated and re-worked by high-flow events, validating the results against in-situ-derived pebble counts. A consumer-grade DJI Phantom drone was used to capture the aerial data used in the SfM workflow. The authors report different relationships between photogrammetry-derived, TLS-derived and in situ observations of grain-size distributions, due to the varying geometry of the SfM acquisitions compared to the ground-based oblique-viewing TLS. They also report a bias in the SfM workflow towards the sampling of coarser particles, compared to classical in situ methods, but conclude by evidencing the various ways in which SfM photogrammetric methods offer other advantages over TLS for such applications (including the “ability to parameterise entrainment and transport rate models for gravel beds by quantifying texture and structure characteristics”).

VI Conclusion

SfM photogrammetry approaches have evolved rapidly over the past two decades. At the end of this special issue, Fawcett et al. revisit a “classic” paper by Chandler, published in 1999, which was amongst the first to discuss digital photogrammetry approaches and their potential application within geospatial sciences. Fawcett et al. discuss how many of Chandler’s early recommendations and insights remain highly relevant within contemporary SfM workflows, whilst also demonstrating the extent to which the field of SfM photogrammetry has evolved since 1999. It is, indeed, testament to the rapid evolution of SfM photogrammetry that a paper published as recently as 1999 should be considered a “classic”!

It has been a pleasure to oversee the process of compiling this special issue. The exceptional papers contained herein are testament to the diversity and quality of scientific work now being undertaken across the environmental, ecological and geosciences disciplines utilising SfM photogrammetry. We believe that the papers contained herein evidence the establishment of SfM photogrammetry as an operational approach to allow user-derived, quantitative, volumetric (and in some cases, uncertainty-assessed) data to be captured using low-cost sensors. We are not advocates of the view that SfM photogrammetry will replace other fine-grained surveying approaches (e.g. terrestrial and ALS) by the virtue of SfM delivering a different type of data to such approaches. However, the upsurge in SfM photogrammetry’s use within the geosciences over the past decades (see Figure 1) is a sign that these techniques will continue to deliver data that complement a wide range of other surveying approaches. We look forward to seeing where SfM photogrammetry takes the disciplines of physical geography and environmental science in the future. Also in this issue Scholefield et al investigate how SfM-derived topographic products combined with aerial orthomosaic-derived habitat classes can be used to derive carbon loss estimates from erosion features in remote blanket peatland habitats. Using a random forest classifier, they show the capabilities of the approach to discriminate exposed peat from saturated bog and sphagnum-dominated habitats. Their method is relatively straightforward to execute, requiring just a lightweight drone equipped with a consumer-grade camera, and accompanied by basic ground control, and yet the method allows for detailed assessment of peatland systems which are otherwise difficult to survey by virtue of their remote locations and inaccessible surface features.

References

Anderson

Gaston

(2013) Lightweight unmanned aerial vehicles will revolutionize spatial ecology. Frontiers in Ecology and the Environment 11: 138–146.

Bakker

Lane

(2017) Archival photogrammetric analysis of river–floodplain systems using Structure from Motion (SfM) methods. Earth Surface Processes and Landforms 42: 1274–1286.

Balsi

Esposito

Fallavollita

. (2018) Single-tree detection in high-density LiDAR data from UAV-based survey. European Journal of Remote Sensing 51: 679–692.

Bayley

DTI

Mogg

AOM

Koldewey

. (2019) Capturing complexity: field-testing the use of ‘structure from motion’ derived virtual models to replicate standard measures of reef physical structure. PeerJ 7: e6540. DOI: https://doi.org/10.7717/peerj.6540.

Bemis

Micklethwaite

Turner

. (2014) Ground-based and UAV-based photogrammetry: A multi-scale, high-resolution mapping tool for structural geology and paleoseismology. Journal of Structural Geology 69: 163–178.

Birdal

Avdan

Türk

(2017) Estimating tree heights with images from an unmanned aerial vehicle. Geomatics, Natural Hazards and Risk 8: 1144–1156.

Brunier

Fleury

Anthony

. (2016) Close-range airborne structure-from-motion photogrammetry for high-resolution beach morphometric surveys: Examples from an embayed rotating beach. Geomorphology 261: 76–88.

Burns

JHR

Delparte

Gates

. (2015) Integrating structure-from-motion photogrammetry with geospatial software as a novel technique for quantifying 3D ecological characteristics of coral reefs. PeerJ 3: e1077.

Carbonneau

Dietrich

(2017) Cost-effective non-metric photogrammetry from consumer-grade sUAS: Implications for direct georeferencing of structure from motion photogrammetry. Earth Surface Processes and Landforms 42: 473–486.

10.

Casella

Collin

Harris

. 2017. Mapping coral reefs using consumer-grade drones and structure from motion photogrammetry techniques. Coral Reefs 36: 269–275.

11.

Castillo

Pérez

James

. (2012) Comparing the accuracy of several field methods for measuring gully erosion. Soil Science Society of America Journal 76: 1319–1332.

12.

Cullen

Verma

Bourke

(2018) A comparison of structure from motion photogrammetry and the traversing micro-erosion meter for measuring erosion on shore platforms. Earth Surface Dynamics 6: 1023–1039.

13.

Cunliffe

Anderson

Brazier

(2016) Ultra-fine grain landscape-scale monitoring of dryland vegetation structure with drone-acquired structure-from-motion SfM photogrammetry. Remote Sensing of Environment 183: 129–143.

14.

Dall’Asta

Thoeni

Santise

. (2015) Network design and quality checks in automatic orientation of close-range photogrammetric blocks. Sensors 15: 7985–8008.

15.

Dandois

Ellis

(2010) Remote sensing of vegetation structure using computer vision. Remote Sensing 2: 1157–1176.

16.

Dietrich

(2017) Bathymetric structure-from-motion: Extracting shallow stream bathymetry from multi-view stereo photogrammetry. Earth Surface Processes and Landforms 42: 355–364.

17.

Dowling

Read

Gallant

(2009) Very high resolution DEM acquisition at low cost using a digital camera and free software. In: 18th world IMACS congress and MODSIM09 international congress on modelling and simulation (eds. Anderssen

Braddock

Newham

LTH

), 2009, 2479–2485. Modelling and Simulation Society of Australia and New Zealand and International Association for Mathematics and Computers in Simulation.

18.

Duffy

Shutler

Witt

. (2018) Tracking fine-scale structural changes in coastal dune morphology using kite aerial photography and uncertainty-assessed structure-from-motion photogrammetry. Remote Sensing 10: 1494.

19.

Eltner

Kaiser

Castillo

. 2016. Image-based surface reconstruction in geomorphometry–merits, limits and developments. Earth Surface Dynamics 4: 359–389.

20.

Fonstad

Dietrich

Courville

. (2013) Topographic structure from motion: A new development in photogrammetric measurement. Earth Surface Processes and Landforms 38: 421–430.

21.

Forsmoo

Anderson

Macleod

CJA

. (2018) Drone-based structure-from-motion photogrammetry captures grassland sward height variability. Journal of Applied Ecology 55: 2587–2599.

22.

Garrett

Anderson

(2018) Drone methodologies: Taking flight in human and physical geography. Transactions of the Institute of British Geographers 43: 341–359.

23.

Glendell

Mcshane

Farrow

. (2017) Testing the utility of structure-from-motion photogrammetry reconstructions using small unmanned aerial vehicles and ground photography to estimate the extent of upland soil erosion. Earth Surface Processes and Landforms 42: 1860–1871.

24.

Holman

Riche

Michalski

. (2016) High throughput field phenotyping of wheat plant height and growth rate in field plot trials using UAV based remote sensing. Remote Sensing 8: 1031.

25.

Immerzeel

Kraaijenbrink

PDA

Shea

. (2014) High-resolution monitoring of Himalayan glacier dynamics using unmanned aerial vehicles. Remote Sensing of Environment 150: 93–103.

26.

James

Robson

(2012) Straightforward reconstruction of 3D surfaces and topography with a camera: Accuracy and geoscience application. Journal of Geophysical Research: Earth Surface 117(F3). DOI: 10.1029/2011JF002289.

27.

James

Applegarth

Pinkerton

(2012) Lava channel roofing, overflows, breaches and switching: Insights from the 2008–2009 eruption of Mt. Etna. Bulletin of Volcanology 74: 107–117.

28.

James

Robson

d’Oleire-Oltmanns

. (2017a) Optimising UAV topographic surveys processed with structure-from-motion: Ground control quality, quantity and bundle adjustment. Geomorphology 280: 51–66.

29.

James

Robson

Smith

(2017b) 3-D uncertainty-based topographic change detection with structure-from-motion photogrammetry: Precision maps for ground control and directly georeferenced surveys. Earth Surface Processes and Landforms 42: 1769–1788.

30.

Javernick

Brasington

Caruso

(2014) Modeling the topography of shallow braided rivers using structure-from-motion photogrammetry. Geomorphology 213: 166–182.

31.

Kääb

Girod

LMR

Berthling

(2014) Surface kinematics of periglacial sorted circles using structure-from-motion technology. The Cryosphere 8: 1041–1056.

32.

Kalacska

Chmura

Lucanus

. (2017) Structure from motion will revolutionize analyses of tidal wetland landscapes. Remote Sensing of Environment 199: 14–24.

33.

Leon

Roelfsema

Saunders

. (2015) Measuring coral reef terrain roughness using ‘Structure-from-Motion’ close-range photogrammetry. Geomorphology 242: 21–28.

34.

Mallalieu

Carrivick

Quincey

. (2017) An integrated structure-from-motion and time-lapse technique for quantifying ice-margin dynamics. Journal of Glaciology 63: 937–949.

35.

Mancini

Dubbini

Gattelli

. (2013) Using Unmanned Aerial Vehicles (UAV) for high-resolution reconstruction of topography: The structure from motion approach on coastal environments. Remote Sensing 5: 6880–6898.

36.

Marteau

Vericat

Gibbins

. (2017) Application of structure-from-motion photogrammetry to river restoration. Earth Surface Processes and Landforms 42: 503–515.

37.

Mercer

Westbrook

(2016) Ultrahigh-resolution mapping of peatland microform using ground-based structure from motion with multiview stereo. Journal of Geophysical Research: Biogeosciences 121: 2901–2916.

38.

Messinger

Asner

Silman

(2016) Rapid assessments of Amazon forest structure and biomass using small unmanned aerial systems. Remote Sensing 8: 615.

39.

Midgley

Tonkin

(2017) Reconstruction of former glacier surface topography from archive oblique aerial images. Geomorphology 282: 18–26.

40.

Mikita

Janata

Surový

(2016) Forest stand inventory based on combined aerial and terrestrial close-range photogrammetry. Forests 7: 165.

41.

Mlambo

Woodhouse

Gerard

. (2017) Structure from Motion (SfM) photogrammetry with drone data: A low cost method for monitoring greenhouse gas emissions from forests in developing countries. Forests 8: 68.

42.

Mölg

Bolch

(2017) Structure-from-motion using historical aerial images to analyse changes in glacier surface elevation. Remote Sensing 9: 1021.

43.

Morgan

Brogan

Nelson

(2017) Application of structure-from-motion photogrammetry in laboratory flumes. Geomorphology 276: 125–143.

44.

Murtiyoso

Grussenmeyer

Börlin

. (2018) Open source and independent methods for bundle adjustment assessment in close-range UAV photogrammetry. Drones 2: 3.

45.

Nakano

Kamiya

Tobita

. (2014) Landform monitoring in active volcano by UAV and SFM-MVS technique. The International Archives of Photogrammetry, Remote Sensing and Spatial Information Sciences 40: 71.

46.

Niethammer

Rothmund

James

. (2010) UAV-based remote sensing of landslides. International Archives of Photogrammetry, Remote Sensing and Spatial Information Sciences 38: 496–501.

47.

Nolan

Larsen

Sturm

(2015) Mapping snow-depth from manned-aircraft on landscape scales at centimeter resolution using structure-from-motion photogrammetry. Cryosphere Discussions 9: 1445–1463.

48.

O’Connor

Smith

James

(2017) Cameras and settings for aerial surveys in the geosciences: Optimising image data. Progress in Physical Geography: Earth and Environment 41: 325–344.

49.

Ozdarici-Ok

(2018) 2-D delineation of individual citrus trees from UAV-based dense photogrammetric surface models. International Journal of Digital Earth 11: 583–608.

50.

Olsoy

Shipley

Rachlow

. (2018) Unmanned aerial systems measure structural habitat features for wildlife across multiple scales. Methods in Ecology and Evolution 9: 594–604.

51.

Prosdocimi

Burguet

Di Prima

. (2017) Rainfall simulation and structure-from-motion photogrammetry for the analysis of soil water erosion in Mediterranean vineyards. Science of the Total Environment 574: 204–215.

52.

Prosdocimi

Calligaro

Sofia

. (2015) Bank erosion in agricultural drainage networks: New challenges from structure-from-motion photogrammetry for post-event analysis. Earth Surface Processes and Landforms 40: 1891–1906.

53.

Rosnell

Honkavaara

(2012) Point cloud generation from aerial image data acquired by a quadrocopter type micro unmanned aerial vehicle and a digital still camera. Sensors 12: 453–480.

54.

Ružić

Marović

Benac

. (2014) Coastal cliff geometry derived from structure-from-motion photogrammetry at Stara Baška, Krk Island, Croatia. Geo-Marine Letters 34: 555–565.

55.

Ryan

Hubbard

Box

. (2015) UAV photogrammetry and structure from motion to assess calving dynamics at Store Glacier, a large outlet draining the Greenland ice sheet. The Cryosphere 9: 1–11.

56.

Smith

Vericat

(2015) From experimental plots to experimental landscapes: Topography, erosion and deposition in sub-humid badlands from structure-from-motion photogrammetry. Earth Surface Processes and Landforms 40: 1656–1671.

57.

Smith

Carrivick

Hooke

. (2014) Reconstructing flash flood magnitudes using ‘Structure-from-Motion’: A rapid assessment tool. Journal of Hydrology 519: 1914–1927.

58.

Stimpson

Gertisser

Montenari

. (2010) Multi-scale geological outcrop visualisation: Using Gigapan and Photosynth in fieldwork-related geology teaching. EGU General Assembly Conference Abstracts. 4702.

59.

Storlazzi

Dartnell

Hatcher

. (2016) End of the chain? Rugosity and fine-scale bathymetry from existing underwater digital imagery using structure-from-motion (SfM) technology. Coral Reefs 35: 889–894.

60.

Turner

Lucieer

Watson

(2012) An automated technique for generating georectified mosaics from ultra-high resolution Unmanned Aerial Vehicle (UAV) imagery, based on Structure from Motion (SfM) point clouds. Remote Sensing 4: 1392–1410.

61.

Varela

Patrício

Anderson

. (2019) Assessing climate change associated sea-level rise impacts on sea turtle nesting beaches using drones, photogrammetry and a novel GPS system. Global Change Biology 25(2): 753–762.

62.

Vázquez-Tarrío

Borgniet

Liébault

. (2017) Using UAS optical imagery and SfM photogrammetry to characterize the surface grain size of gravel bars in a braided river (Vénéon River, French Alps). Geomorphology 285: 94–105.

63.

Wallace

Lucieer

Malenovsky

. (2016) Assessment of forest structure using two UAV techniques: A comparison of airborne laser scanning and Structure from Motion (SfM) point clouds. Forests 7: 62.

64.

Webster

Westoby

Rutter

. (2018) Three-dimensional thermal characterization of forest canopies using UAV photogrammetry. Remote Sensing of Environment 209: 835–847.

65.

Welty

Pfeffer

Ahn

(2010) Something for everyone: Quantifying evolving (glacial) landscapes with your camera. AGU Fall Meeting Abstracts.

66.

Westoby

Brasington

Glasser

. (2012) ‘Structure-from-Motion’ photogrammetry: A low-cost, effective tool for geoscience applications. Geomorphology 179: 300–314.

67.

Woodget

Carbonneau

Visser

. (2015) Quantifying submerged fluvial topography using hyperspatial resolution UAS imagery and structure from motion photogrammetry. Earth Surface Processes and Landforms 40: 47–64.

68.

Zahawi

Dandois

Holl

. (2015) Using lightweight unmanned aerial vehicles to monitor tropical forest recovery. Biological Conservation 186: 287–295.