Table 1.
Comparison of the altimetry and stereo-video camera methods for morphometric measurements.
Fig 1.
Examining the feasibility of using stereo-video footage collected from the air to measure objects below the water surface.
(A) Calibrating the SeaGIS SVC to a custom scale bar before testing. (B) SeaGIS swimmable stereo-video camera on the high dive, stereo-video camera in red dashed outline. (C) Underwater scale bar testing mechanism at 5.0 m depth. Testing was conducted at the University of West Florida Aquatic Center.
Fig 2.
Conceptual overview of stereo-video camera mounting to an unoccupied aerial vehicle.
Dashed lines indicate the convergence points of left and right cameras on object, relative to UAV height above the water surface and underwater object. (Turtle silhouette image by Spotila and Chatterji—no copyright).
Fig 3.
Unoccupied aerial vehicle in flight with stereo-video camera system mounted, and left camera highlighted.
Fig 4.
Custom ABS plastic 3D printed housing, which allows access to camera functions for the GoPro Hero 5 Black camera. Housing is fastened using stainless steel screws and wingnuts for corrosion resistance.
Fig 5.
Stereo-video camera software interface.
(A) Left and (B) right camera views of a nurse shark (Ginglymostoma cirratum) measured for total length, and (C) zoomed-in image of the right camera image of the shark. In (C), the photo has been enhanced for contrast and to sharpen edges to aid in shark visibility. Red lines indicate the measurement vectors for total length.
Fig 6.
Relationship of absolute error of aerial SVC measurements and depth.
The solid line and gray shaded area are the generalized linear regression and standard error, respectively. Depth measurements of <0 indicate measurements made from the camera in the air only, just above the water surface.
Table 2.
Measurement accuracy of the stereo-video camera mounted to an unoccupied aerial vehicle and in-water swimmable stereo-video camera.
Fig 7.
Comparison of hand measurements to SVC-UAV for green sea turtles (n = 6) from the field trial in Eleuthera, Bahamas.
The dashed line and gray shaded area are the linear regression and standard error, respectively. The solid black line is the 1:1 line for the hand and SV-UAV measurements; points that fall on this line have equivalent measurements for both methods.
Fig 8.
Relationship between absolute error and (A) range and (B) straight carapace length from the stereo-video camera system mounted to an unoccupied aerial vehicle.
The dashed line and gray shaded area are the linear regression and standard error, respectively.