Page 4 - i1052-5173-31-9

Virtual Outcrops in a Pocket:

The Smartphone as a Fully Equipped

Photogrammetric Data Acquisition Tool

Amerigo Corradetti*, Dept. of Mathematics and Geosciences, University of Trieste, Trieste, Italy; Thomas D. Seers, Dept. of Petroleum
Engineering, Texas A&M University at Qatar, Doha, Qatar; Andrea Billi, Consiglio Nazionale delle Ricerche, IGAG, Rome, Italy; and
Stefano Tavani, Consiglio Nazionale delle Ricerche, IGAG, Rome, Italy, and DiSTAR, Università di Napoli Federico II, Napoli, Italy

ABSTRACT seismically active fault. The scan is con- grammetry by outcrop geologists was ini-
Since the advent of affordable consumer- ducted with minimal effort over the course tially slow (e.g., Hodgetts et al., 2004; Pringle
grade cameras over a century ago, photo- of a few minutes with limited equipment, et al., 2004), with legacy photogrammetric
graphic images have been the standard thus being representative of a routine situa- reconstruction techniques requiring highly
medium for capturing and visualizing out- tion for a field geologist. specialized, expensive metric cameras or
crop-scale geological features. Despite the software (Chandler and Fryer, 2005), and
ubiquity of raster image data capture in rou- INTRODUCTION AND commonly carried the limitation of cumber-
tine fieldwork, the development of close- BACKGROUND some manual assignment of key points on
range 3D remote-sensing techniques has led Rapid improvements in the fidelity of con- the targeted rock surface (e.g., Simpson et al.,
to a paradigm shift in the representation and sumer-grade cameras, coupled with novel 2004). Many of these disadvantages were
analysis of rock exposures from two- to computer vision–based photogrammetric addressed with the advent of low-cost or
three-dimensional forms. The use of geolog- image processing pipelines (i.e., structure open-source SfM-MVS photogrammetry
ical 3D surface reconstructions in routine from motion–multiview stereo photogram- image processing pipelines (e.g., Snavely et
fieldwork has, however, been limited by the metry: SfM-MVS), have revolutionized out- al., 2006; Furukawa and Ponce, 2009; Wu,
portability, associated learning curve, and/ crop studies over the past decade, bringing 2011), which facilitated the use of uncali-
or expense of tools required for data capture, traditional field geology into the digital age. brated consumer-grade cameras and enabled
visualization, and analysis. Smartphones These developments are also closely tied to automated image key-point detection and
are rapidly becoming a viable alternative to major methodological improvements for vir- matching (e.g., James and Robson, 2012).
conventional 3D close-range remote-sensing tual outcrop model (VOM) interpretation. The potential of producing 3D rock-surface
data capture and visualization platforms, All these advancements have accelerated the models using consumer-grade cameras
providing a catalyst for the general uptake of use of digital outcrop data capture and analy- attracted the interest of numerous workers.
3D outcrop technologies by the geological sis in field geology, transforming what was These developments coupled with the increas-
community, which were up until relatively principally a visualization medium into fully ing availability of lightweight and low-cost
recently the purview of a relatively small interrogatable quantitative geo-data objects drones able to carry cameras and other sen-
number of geospatial specialists. Indeed, the (Jones et al., 2004; Bemis et al., 2014; Howell sors, have finally boosted the use of SfM-
continuous improvement of smartphone et al., 2014; Hodgetts et al., 2015; Biber et al., MVS reconstruction in geosciences.
cameras, coupled with their integration with 2018; Bruna et al., 2019; Caravaca et al., For many geoscience applications, it is
global navigation satellite system (GNSS) 2019; Thiele et al., 2019; Triantafyllou et al., necessary to register 3D rock-surface recon-
and inertial sensors provides 3D reconstruc- 2019). Initially, close-range remote-sensing structions within a local or global coordinate
tions with comparable accuracy to survey- studies seeking to reconstruct and analyze frame. The use of survey-grade total stations
grade systems. These developments have rock outcrops were dominantly built around and/or real-time kinematic (RTK) differen-
already led many field geologists to replace terrestrial laser scanning systems (terrestrial tial global navigation satellite system (GNSS)
reflex cameras, as well as dedicated hand- lidar), which became commercially available antennas permit both terrestrial (Jaud et al.,
held GNSS receivers and compass clinome- around two decades ago (e.g., Bellian et al., 2020) and aerial (Rieke et al., 2012) image
ters, with smartphones, which offer the 2002). These initial works tended to be tech- data and/or ground control points (GCPs) to
equivalent functionality within a single nology demonstrations rather than routine be georeferenced within the mapped scene
compact platform. Here we demonstrate that field studies, with the expense, weight, and with centimeter to millimeter accuracy
through the use of a smartphone and a por- challenging operational learning curve lim- (Bemis et al., 2014). Those survey tools are,
table gimbal stabilizer, we can readily gen- iting replication to a few highly specialized however, bulky and expensive, and are not
erate and register high-quality 3D scans of geospatial specialists and groups. Receiving standard tools for geoscientists engaged in
outcropping geological structures, with the greater interest from the archaeological fieldwork. Improvements in consumer-grade
workflow exemplified using a mirror of a community, the adoption of digital photo- GNSS receivers, capable of harnessing

GSA Today, v. 31, https://doi.org/10.1130/GSATG506A.1. CC-BY-NC.

*amerigo.corradetti@units.it

4 GSA Today | September 2021