Meet the Bass Connections Team
Meet the Team:
We are the 2018 Bass Connections team, working with the Marine Robotics and Remote Sensing Lab. We’re excited to share our research path with you as we develop methods to evaluate coastal habitats using unoccupied aerial systems, or drones.
Our team is led by postdoctoral researcher Dr. Justin Ridge, who received his Ph.D. in Marine Sciences from UNC IMS. Our Project Manager is Master of Environmental Management student Kelly Dobroski, studying Coastal Environmental Management. Before starting her MEM, she worked in the Chesapeake Bay Region as an informal science educator and researcher.
Our second MEM-CEM student researcher is Claire Atkins-Davis. Claire’s interests are in using technology to understand marine mammal acoustics and their environment.
Our undergraduate researchers are Alexandra DiGiacomo and Virginia Pan – both rising juniors at Duke. Alexandra is studying Biology, with a focus in Marine Biology; Virginia is studying Electrical and Computer Engineering and Marine Science. Both undergraduates are Rachel Carson Scholars, a program for select undergraduates to gain first-hand experience with marine science research and conservation.
Our Project Goals:
Living shorelines are critical as they provide many ecosystem services, such as habitat, water filtration, and buffer coastlines from storms. Additionally, they provide the first line of defense in the wake of climate change and rising sea levels; therefore monitoring their change is critical. Current methods are highly invasive, requiring boots on the ground and are time and resource intensive. Our team is working to determine whether drones can be used to accurately conduct living shoreline assessments.
The first phase of fieldwork has begun for the Bass Connections Team. Our overall objective is to assess the accuracy of drones or unoccupied aerial vehicles (UAVs) to be able to measure and monitor “living shoreline” ecosystems. UAVs have the ability to cover larger areas in shorter periods of time, monitor
areas with remote or limited access using field verification methods, as well as impose less disturbance to the ecosystem during monitoring and maintenance. We have worked primarily with salt marshes so far this summer, and look forward to exploring other ecosystems soon!
We have used ground truthing methods concurrently with UAV sensors, allowing us to determine whether or not drones can accurately capture measurements of these coastal habitats. We have spent the first few weeks up to our ears in marsh grasses, counting stem density (the number of plants in a given area) and stem height, both indicators of how healthy a salt marsh is. We use the EMLID REACH RTK
GPS unit to collect elevation measurements and GPS locations to align the aerial imagery with our ground control points. These points are black and white squares that are easily identifiable by the drone in marsh grasses.
Currently, we have surveyed local areas around the Duke Marine Lab Campus and have utilized two different drones and sensors to be able to capture all the variables of interest that allow the Bass Connection team to efficiently monitor the health of these habitats. We recently returned from a two day field experience in Shallotte, NC, where we also conducted these same ground truthing methods on taller, more dense grasses. After data collection, we process the imagery using software such as Pix4D, ArcGIS Pro, and R-Studio.
To analyze the imagery that we receive from the drones, we run many images through a processing software. High overlap of these images allows us to create a 3-D model of our study area in Pix4D. In the past few weeks, we have done two different flights over a salt marsh on DUML’s campus. We have used consumer-grade drones and commercial-grade drones equipped with sensors of different resolution in order to determine which sensors are most effective in collecting living shoreline data.
The 3-D model of the marsh created from Pix4D is represented by millions of points with latitude, longitude and height values. To do this, we first create a 3-D image, or an orthomosaic, in Pix4D. From there, the millions of points generated form a point cloud. From this point cloud, we have used ArcGIS Pro to extract information like height, density, and biomass. This extracted data can be compared to our ground–truthed data, to determine how accurately the drone captures these measurements.
So far, we are having success! We look forward to testing our workflow on different marshes to determine how applicable our approach is. Stay tuned to learn more about our project updates.