Article and Figures Provided By: Tim Battista (NOS/NCCOS/Marine Spatial Ecology Division)
Why We Care
Several organizations, including NOAA, need imagery, elevation and depth data to inform management decisions in coastal areas (Figure 1). Coastal planning and management organizations need information because the United States’ shorelines are busy and dynamic places, encompassing a wide range of ecologies, geologies, hazards, human uses and political jurisdictions. Although there is a great need for data in coastal areas, many near-shore areas are expensive, challenging and even dangerous to map with existing, manned technologies or traditional hydrographic techniques (e.g. airborne LIDAR, or multibeam from launches or ships) resulting in information gaps along the coastline. These regions are very energetic (i.e. surf). Small unmanned aerial systems (sUAS) may offer a potentially safe and effective way to fill these critical gaps, since they are capable of producing centimeter-scale resolution photographs in near-shore environments (Figure 2).
Elevations and depths can be derived from these photographs using a technique known as Structure from Motion (SfM). SfM uses overlapping photographs (collected by the sUAS) and photogrammetry to calculate the heights and depths along the coastline. Structure from Motion (SfM) is a relatively new type of photogrammetry, which leverages advanced computer vision algorithms to enable highly automated processing and the use of non-mapping-grade cameras, such as those typically installed on UAS. Like conventional stereo photogrammetry, SfM relies on sets of overlapping images to reconstruct 3D geometry from 2D imagery. However, unlike conventional photogrammetry, which came into maturity well before today’s advanced computer hardware and software, SfM overcomes the need for highly-calibrated and stable metric mapping cameras and stable imaging geometry through advanced algorithms. Because SfM works well with UAS imagery, is highly automated, and enables generation of elevations with spatial resolutions and accuracies generally comparable to LIDAR, as well as orthophotos. Combining sUAS and SfM could potentially fill nearshore data and information gaps accurately, safely and cost-effectively, providing comprehensive and timely information to coastal managers and planners. However, additional research is needed to identify optimal payloads (i.e. sensors) and processing workflows (i.e. techniques, algorithms, processing), as well as define environmental and operational limitations of these systems before being implemented more widely across NOAA.
What We Are Doing
National Centers for Coastal Ocean Science (NCCOS) scientists are using sUAS and SfM to map nearshore depths and elevations. This work funded by the UAS Program Office is being conducted in close collaboration with Oregon State University (OSU), NOAA’s Office of Coast Survey (OCS), NOAA’s National Geodetic Survey (NGS), and in coordination with local partners including USVI Territorial Government, National Park Service (NPS), NOAA’s Office of National Marine Sanctuaries (ONMS), The Nature Conservancy (TNC), and University of California Santa Barbara (UCSB). SUAS and SfM were first tested in St. Croix in the U.S. Virgin Islands to identify optimal sUAS payloads and SfM processing workflows. These tests occurred in two marine managed areas, including Buck Island Reef National Monument (BIRNM) and the St. Croix East End Marine Park (EEMP). Initial results were promising, leading to follow on research on Santa Cruz Island, California. The objective of this follow on project was to understand the operational and environmental limitations of sUAS and SfM to map coastal elevations and depths in more remote locations and challenging environmental conditions. Tests occurred on land owned by TNC, and on submerged lands under the jurisdiction of Channel Islands National Marine Sanctuary. An Autonomous Surface Vehicle (ASV) was also tested along the Santa Cruz Island coast. Depths were measured by a Sound Navigation and Ranging (SONAR) sensor mounted on the ASV. These depth measurements were used to independently validate the depths derived from SfM, and to better understand the benefit of these surface vehicles for nearshore bathymetric mapping. The lessons learned and results from this research will be compiled in a technical report. This report will help guide future coastal mapping using sUAS and SfM at NOAA, and is designed to be a companion to NOAA Office of Coast Survey’s procedures manual for operating sUAS from NOAA hydrographic ships. The intent of combining these procedures is to help move this technological approach from research and one-step closer to operations at NOAA.
Impact/Benefits of Our Work
The use of commercially available unmanned vehicles have become increasingly common in coastal areas. This work was funded by NOAA’s Unmanned Aircraft Systems Program Office (UASPO) through its federal funding opportunity. Our research was designed to support NOAA OCS’s nautical charting needs in shallow (<10 m) waters. NOAA OCS is developing procedures to operate sUAS from hydrographic vessels. The methods described here will be companion to these operational procedures, and together will help move this technological approach from research towards operations at NOAA (Figure 3). In addition working to meet NOAA OCS’s needs, there has been substantial interest in sUAS applications beyond nautical charting from other NOAA offices, federal agencies, state agencies and non-governmental organizations. This interest has ranged widely from mapping and monitoring coastal habitats to surveying marine animals to observing human activities in coastal environments. Efforts are currently underway to build on the research described here, and continue to design and test new applications for commercially available unmanned vehicles at NOAA NCCOS.
