Search

Structure-from-motion (SfM) photogrammetry techniques are now widely available to generate digital terrain models (DTMs) from optical imagery, providing an alternative to costlier options such as LiDAR or satellite surveys. SfM could be a useful tool in hazard studies because its minimal cost makes it accessible even in developing regions and its speed of use can provide updated data rapidly in hazard-prone regions. Our study is designed to assess whether crowd-sourced SfM data is comparable to an industry standard LiDAR dataset, demonstrating potential real-world use of SfM if employed for disaster risk reduction purposes. Three groups with variable SfM knowledge utilized 16 different camera models, including four camera phones, to collect 1001 total photos in one hour of data collection. Datasets collected by each group were processed using VisualSFM, and the point densities, accuracies and distributions of points in the resultant point clouds (DTM skeletons) were compared. Our results show that the point clouds are resilient to inconsistency in users' SfM knowledge: crowd-sourced data collected by a moderately informed general public yields topography results comparable in data density and accuracy to those produced with data collected by highly-informed SfM users or experts using LiDAR. This means that in a real-world scenario involving participants with a diverse range of expertise, topography models could be produced from crowd-sourced data quite rapidly and to a very high standard. This could be beneficial to disaster risk reduction as a relatively quick, simple and low-cost method to attain rapidly updated knowledge of terrain attributes, useful for the prediction and mitigation of many natural hazards. [ABSTRACT FROM AUTHOR]

Digital photogrammetry has experienced rapid development regarding the technology involved and its ease of use over the past two decades. We revisit the work of Jim Chandler who in 1999 published a technical communication seeking to familiarise novice users of photogrammetric methods with important theoretical concepts and practical considerations. In doing so, we assess considerations such as camera calibration and the need for photo-control and check points, as they apply to modern software and workflows, in particular for structure-from-motion (SfM) photogrammetry. We also highlight the implications of lightweight drones being the new platform of choice for many photogrammetry-based studies in the geosciences. Finally, we present three examples based on our own work, showing the opportunities that SfM photogrammetry offers at different scales and systems: at the micro-scale for monitoring geomorphological change, and at the meso-scale for hydrological modelling and the reconstruction of vegetation canopies. Our examples showcase developments and applications of photogrammetry which go beyond what was considered feasible 20 years ago and indicate future directions that applications may take. Nevertheless, we demonstrate that, in-line with Chandler's recommendations, the pre-calibration of consumer-grade cameras, instead of relying entirely on self-calibration by software, can yield palpable benefits in micro-scale applications and that measurements of sufficient control points are still central to generating reproducible, high-accuracy products. With the unprecedented ease of use and wide areas of application, scientists applying photogrammetric methods would do well to remember basic considerations and seek methods for the validation of generated products. [ABSTRACT FROM AUTHOR]

• UAV-SfM soil loss measurements were validated by direct sediment collection. • Photogrammetric precision estimates defined spatially variable levels of detection. • DEMs of difference illustrated the diffuse erosion patterns through time. • Accurate 3-D soil surface models were constructed from UAV-SfM data. Sheet erosion is common on agricultural lands, and understanding the dynamics of the erosive process as well as the quantification of soil loss is important for both soil scientists and managers. However, measuring rates of soil loss from sheet erosion has proved difficult due to requiring the detection of relatively small surface changes over extended areas. Consequently, such measurements have relied on the use of erosion plots, which have limited spatial coverage and have high operating costs. For measuring the larger erosion rates characteristic of rill and gully erosion, structure-from-motion (SfM) photogrammetry has been demonstrated to be a valuable tool. Here, we demonstrate the first direct validation of UAV-SfM measurements of sheet erosion using sediment collection data collected from erosion plots. Three erosion plots (12 m × 4 m) located at Lavras, Brazil, with bare soil exposed to natural rainfall from which event sediment and runoff was monitored, were mapped during two hydrological years (2016 and 2017), using a UAV equipped with a RGB camera. DEMs of difference (DoD) were calculated to detect spatial changes in the soil surface topography over time and to quantify the volumes of sediments lost or gained. Precision maps were generated to enable precision estimates for both DEMs to be propagated into the DoD as spatially variable vertical uncertainties. The point clouds generated from SfM gave mean errors of ~2.4 mm horizontally (xy) and ~1.9 mm vertically (z) on control and independent check points, and the level of detection (LoD) along the plots ranged from 1.4 mm to 7.4 mm. The soil loss values obtained by SfM were significantly (p < 0.001) correlated (r2 = 95.55%) with those derived from the sediment collection. These results open up the possibility to use SfM for erosion studies where channelized erosion is not the principal mechanism, offering a cost-effective method for gaining new insights into sheet, and interrill, erosion processes. [ABSTRACT FROM AUTHOR]

The recent growth in popularity of photogrammetry for geoscientific studies has largely been driven by the development of Structure from Motion (SfM) algorithms. Their use require relatively low levels of technical expertise for the production of three dimensional point data that can subsequently be used as a primary data source. While much research has gone into optimising camera network design and parameterisation of the SfM process, little attention has been given factors associated with image collection and processing which will impact the image data. This research details the effect of image compression and, specifically, the impact of images acquired in the standard JPEG format with varying levels of compression on the SfM workflow. JPEG compression is categorized by how aggressively portions of the image are quantized by lookup tables using a metric known as 'quality'.This research investigates the impact of image compression within an SfM photogrammetry workflow, quantitatively assessing photogrammetric outputs. To do this, two field surveys were undertaken in Norfolk, UK, where over 5000 images were captured using a Nikon D700 camera and Nikkor 35 mm lens to test a range of image compression scenarios on photogrammetric outputs. These were recorded in the Nikon NEF RAW file format before lossless compression to 16-bit uncompressed TIFF files. Contemporaneously, a control network was surveyed using a Leica total station data (TCR-805; estimated 3 mm distance accuracy) and a further point cloud was captured with a Leica terrestrial laser scanner (P40, estimated 3 mm 1-sigma accuracy).We find that JPEG compression degrades photogrammetric results, and should not be used when accuracy is a high priority. For JPEG images produced at low levels of compression (compression 'quality' 92; typical default JPEG compression), an increase of 4.3 % in median cloud-cloud error is observed when compared to the TIFF image files. Point count is also reduced, though only slightly, with the TIFF image files producing 1 % more points. Increasing compression to a 'quality' of 50 (similar to Nikon 'basic image' compression) increases this error to 13.2 % when compared with the TIFF image files. Point count is also decreased, with the clouds containing 7.5 % less points than those produced from the TIFF image files. Whilst these results are scene-dependant, a trend in degradation of photogrammetric products with increased compression is evident across all imagery analysed within this experiment. We conclude that image compression and data handling are important considerations within the SfM workflow and consequently RAW image data should be required for use within photogrammetric workflows where possible. [ABSTRACT FROM AUTHOR]