Utilizing Small Unmanned Aerial Systems (sUAS) with Structure from Motion (SfM) for Shoreline Verification

Abstract

Utilizing Small Unmanned Aerial Systems (UAS) with Structure from Motion (SfM) for Shoreline Verification from NOAA Platforms The National Oceanic and Atmospheric Administration’s (NOAA) hydrographic field parties tasked by the Office of Coast Survey typically conduct limited shoreline verification for surveys that include the near-shore environment. Limited shoreline verification is restricted to the navigable area (typically seaward of the 4 meter contour). In this area, surveyors confirm or disprove the presence of features, where a feature can be any anthropogenic or natural object that may merit individual cartographic representation (e.g., rocks, wrecks,obstructions), identify and position features not detected using aircraft or satellite based photogrammetry, and obtain heights on significant features. Currently, NOAA’s Office of Marine and Aviation Operations field units utilize hydrographic survey launches or skiffs with personnel on board to get close enough to shoreline features to obtain a position, an activity that is generally inefficient and inherently dangerous. Other technologies have been applied to this task with differing degrees of success. Examples include airborne lidar, vessel mounted, side-looking lidar, and satellite based bathymetry. Each of these have each demonstrated certain advantages, but also come with operational and logistical constraints, environmental limitation, and mobilization cost. Advancements in Unmanned Aerial systems (UAS) and photogrammetric techniques, specifically structure from motion (SfM), make these platforms a tempting choice for this mission. The combination of relatively low vehicle and sensor cost, small platform size, and ease of operations may allow collection of high-resolution shoreline data from a safe distance,without ever having to put manned platforms in harm’s way. Under this approach, the surveyor deploys an UAS with a mounted digital camera and flies pre-programmed flight paths to cover a survey area with overlapping aerial photographs. SfM processing of this imagery yields both 3D point clouds and an image mosaic. We report our success to date with this approach, including the substantial regulatory and administrative aspects of planning and executing these missions. While the regulatory aspects of unmanned aerial systems are rapidly evolving, we outline a reasonable and well-vetted approach for operators that has been successful for our work. We believe that demonstrating a consistent and pro-active commitment to safe operations and full compliance with environmental and regulatory aspects will be critical to the development of this technology. We share the operation risk management and best management practices developed so far. We also report on our data processing pipeline and findings from operations conducted in realistic field environments. Similar to other work in this field, we find proper image and environmental parameters are critical. Sun glint, cloud cover, and image quality are important. We also find waves (in particular breaking waves) cause difficulty in SfM processing, but interestingly, often convey important information about submerged features that are of particular interest. Where image processing to a SfM model is successful, we find positional accuracies are well within horizontal product specifications horizontally - even with inexpensive, consumer grade GNSS systems. Precise vertical control is more of a challenge. We report a tradeoff between establishing a network of ground control stations and establishing robust vehicle positioning through more advance methods such as post-processed kinematic PPK). With further development, these UAS derived products may augment traditional shoreline products to provide the high resolution and confidence to eliminate or significantly curtail the need to operate manned boats in the near shore environment.