A pinboard by
this curator

PhD candidate studying the interplay of modern technology and traditional field techniques.


Photogrammetry is an ideal tool for studying the Earth, bringing outcrops from the field to the lab.

In ten seconds? Structure from Motion (SfM) Photogrammetry algorithms allow for the creation of 3D point clouds and surfaces just from overlapping photographs. This is ideal for studying the shape and changes of the world around us.

Don't believe it? Researchers have found uses for photogrammetry which span monitoring volcanoes to mapping sedimentological features.

But how accurate are these models? This can vary based on the number of photographs and ground control accuracy. However, comparisons to more traditional LiDAR techniques show that photogrammetry stacks up very well.

So how is this technology being used?

Integrated with UAV technology - With the advent of consumer UAVs researchers have been able to take their photographs from the air and from vantages previously unavailable or too difficult to obtain. The addition of UAV technology to the SfM workflow also broadens the spatial expanse which can be modeled. A study showed that ground based SfM was much less cost effective in large study areas compared to UAV based SfM and airborne LiDAR.

Use with multi-spectral images - Traditionally SfM is used with visible imagery, however, a study monitoring volcanic activity used SfM to build 3D models of thermal images. The use of multi-spectral imagery in SfM photogrammetry may reveal details of a study area not able to be seen with visible light photography.

Changes in our landscape - While much of geology deals with non-human time scales our understanding of the past comes from understanding how the Earth changes and reacts today. Measuring changes in topography due to erosion or changes in glacial extent are key to understanding our Earth system. Researchers have been able to use archived images to recreate glacial surfaces of the past and compare them to today enhancing our understanding of glacial evolution.


Reconstruction of former glacier surface topography from archive oblique aerial images

Abstract: Archive oblique aerial imagery offers the potential to reconstruct the former geometry of valley glaciers and other landscape surfaces. Whilst the use of Structure-from-Motion (SfM) photogrammetry with multiview stereopsis (MVS) to process small-format imagery is now well established in the geosciences, the potential of the technique for extracting topographic data from archive oblique aerial imagery is unclear. Here, SfM-MVS is used to reconstruct the former topography of two high-Arctic glaciers (Midtre and Austre Lovénbreen, Svalbard, Norway) using three archive oblique aerial images obtained by the Norwegian Polar Institute in 1936. The 1936 point cloud was produced using seven LiDAR-derived ground control points located on stable surfaces in proximity to the former piedmont glacier termini. To assess accuracy, the 1936 data set was compared to a LiDAR data set using the M3C2 algorithm to calculate cloud-to-cloud differences. For stable areas (such as nonglacial surfaces), vertical differences were detected between the two point clouds (RMS M3C2 vertical difference of 8.5 m), with the outwash zones adjacent to the assessed glacier termini showing less extensive vertical discrepancies (94% of M3C2 vertical differences between ± 5 m). This research highlights that historical glacier surface topography can be extracted from archive oblique aerial imagery, but accuracy is limited by issues including the lack of camera calibration, the quality and resolution of the archive imagery, and by the identification of suitable ground control. To demonstrate the value of historical glacier surfaces produced using oblique archive imagery, the reconstructed glacier surface topography is used to investigate evidence of a potential former surge front at the high-Arctic valley glacier, Austre Lovénbreen — a glacier identified to have potentially exhibited surge-type behaviour during the Neoglacial. A surface bulge of ~ 15–20 m is resolved on the 1936 model; however, when compared with the now deglaciated former subglacial topography, a surge origin for the surface feature becomes unclear. The processed 1936 oblique imagery was also used to produce orthorectified nadir aerial imagery, from which structural mapping was undertaken: this adds to the existing 1948–1995 structural map series for these glaciers. This research demonstrates the potential of SfM-MVS for reconstructing historical glacier surfaces, which is important for aiding our understanding of former glacier dynamics and enabling the rapid assessment of glacier change over time.

Pub.: 06 Jan '17, Pinned: 21 Apr '17

3-D uncertainty-based topographic change detection with structure-from-motion photogrammetry: precision maps for ground control and directly georeferenced surveys

Abstract: Structure-from-motion (SfM) photogrammetry is revolutionising the collection of detailed topographic data, but insight into geomorphological processes is currently restricted by our limited understanding of SfM survey uncertainties. Here, we present an approach that, for the first time, specifically accounts for the spatially variable precision inherent to photo-based surveys, and enables confidence-bounded quantification of 3D topographic change. The method uses novel 3D precision maps that describe the 3D photogrammetric and georeferencing uncertainty, and determines change through an adapted state-of-the-art fully 3D point-cloud comparison (M3C2), which is particularly valuable for complex topography. We introduce this method by: (1) using simulated UAV surveys, processed in photogrammetric software, to illustrate the spatial variability of precision and the relative influences of photogrammetric (e.g. image network geometry, tie point quality) and georeferencing (e.g. control measurement) considerations; (2) we then present a new Monte Carlo procedure for deriving this information using standard SfM software and integrate it into confidence-bounded change detection; before (3) demonstrating geomorphological application in which we use benchmark TLS data for validation and then estimate sediment budgets through differencing annual SfM surveys of an eroding badland. We show how 3D precision maps enable more probable erosion patterns to be identified than existing analyses, and how a similar overall survey precision could have been achieved with direct survey georeferencing for camera position data with precision half as good as the GCPs'. Where precision is limited by weak georeferencing (e.g. camera positions with multi-metre precision, such as from a consumer UAV), then overall survey precision can scale as n-½ of the control precision (n = number of images). Our method also provides variance–covariance information for all parameters. Thus, we now open the door for SfM practitioners to use the comprehensive analyses that have underpinned rigorous photogrammetric approaches over the last half-century. Copyright © 2017 John Wiley & Sons, Ltd.

Pub.: 16 Mar '17, Pinned: 21 Apr '17

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

Abstract: Quantifying the extent of soil erosion at a fine spatial resolution can be time consuming and costly; however, proximal remote sensing approaches to collect topographic data present an emerging alternative for quantifying soil volumes lost via erosion. Herein we compare terrestrial laser scanning (TLS), and both unmanned aerial vehicle (UAV) and ground photography (GP) structure-from-motion (SfM) derived topography. We compare the cost-effectiveness and accuracy of both SfM techniques to TLS for erosion gully surveying in upland landscapes, treating TLS as a benchmark. Further, we quantify volumetric soil loss estimates from upland gullies using digital surface models derived by each technique and subtracted from an interpolated pre-erosion surface. Soil loss estimates from UAV and GP SfM reconstructions were comparable to those from TLS, whereby the slopes of the relationship between all three techniques were not significantly different from 1:1 line. Only for the TLS to GP comparison was the intercept significantly different from zero, showing that GP is more capable of measuring the volumes of very small erosion features. In terms of cost-effectiveness in data collection and processing time, both UAV and GP were comparable with the TLS on a per-site basis (13.4 and 8.2 person-hours versus 13.4 for TLS); however, GP was less suitable for surveying larger areas (127 person-hours per ha−1 versus 4.5 for UAV and 3.9 for TLS). Annual repeat surveys using GP were capable of detecting mean vertical erosion change on peaty soils. These first published estimates of whole gully erosion rates (0.077 m a−1) suggest that combined erosion rates on gully floors and walls are around three times the value of previous estimates, which largely characterize wind and rainsplash erosion of gully walls. Copyright © 2017 John Wiley & Sons, Ltd.

Pub.: 10 Apr '17, Pinned: 21 Apr '17