Black and white imagery were collected using a stereo pair of underwater video cameras mounted on a SeaBED autonomous underwater vehicle (AUV) and deployed around the island of Tutuila in American Samoa to record observations of fishes encountered there. Imagery were collected during 17 AUV dives conducted on 10 different days during April 2012.
The stereo video data were one of several different fishery-independent methodologies being cross-compared to evaluate their efficacy as a means of sampling coral reef fish assemblages.
The island of Tutuila is the principal island in American Samoa, home to the largest human population and the seat of government. It features the safest deep water harbor in the central Pacific and by far the largest insular shelf, in American Samoa, averaging perhaps 5 km wide. Operations Description: From April 2 - April 19, 2012, CRED personnel and partners from other divisions in the Pacific Islands Fisheries Science Center and Northwest Fisheries Science Center participated in cruise SE1202. During the cruise a total of 17 AUV dives were conducted around Tutuila, off the NOAA Ship Oscar Elton Sette. Dives were conducted during daylight hours and collectively sampled reef fish communities over 29.4 km of seafloor. The maximum depth reached on any dive was 150 m, but the vast majority of the more than 15 hours of bottom time was spend at depths of 30 - 60 m. Equipment Description: The SeaBED-class AUV, unlike other more traditional AUV's, employs a twin-hull design that provides enhanced stability for low-speed photographic surveys. Designed and built by the lab of Dr. Hanaumant Singh at the Woods Hole Oceanographic Institute (WHOI), SeaBED is designed to autonomously follow the terrain approximately 3 to 4 m above the sea floor, collecting high resolution color and black-and-white imagery while maintaining a forward speed of .25 - .5 m/sec. For this mission, SeaBed was also be outfitted with a forward-looking stereo video camera system as well as a forward-looking imaging SONAR unit. The stereo-video system is similar to that use on the BRUVS and allows for accurate measures of fish abundance and size structure. The imaging SONAR unit is being tested as a means to assess fish assemblage outside the visual range of the cameras and in zero light situations including nocturnal or operations in depths to which light does not reach. SeaBED is approximately two meters long and weighs nearly 200 kg. It has two main pressure housings, a top hull and a bottom hull. The CPU electronics are located in the top hull, and the batteries, cameras, and sensors are located in the bottom hull, and all are connected by wet cabling that is routed through vertical struts. With a maximum depth range of 2,000 m, and maximum single-dive time of 6 - 8 hours, SeaBED can be used to survey habitats ranging from shallow coral reefs to deep groundfish environments. The AUV is programmed while still aboard the ship. Programming parameters include navigational waypoints, speed, altitude to maintain above the seafloor, and frequency of photographs. Once submerged, the AUV does not resurface until the end of its mission. An RD Instruments 1200 kHz Doppler Velocity Log, iXSea Octans Inertial Navigation Unit, and Paroscientific Depth Sensor provide the data necessary for the vehicle's autonomous navigation. The AUV does report its position to the ship periodically in telemetry messages via acoustic MODEM. Additionally USBL tracking shows range and bearing between the ship and AUV during the mission. If any of these telemetry messages indicate an unexpected change in the AUV's planned mission, the mission can be aborted via acoustic MODEM message, resulting in the AUV returning to the surface for recovery. The SeaBED AUV carries a forward-facing ROS Navigator black-and-white, low-light stereo-video camera system, two 5 megapixel, 12 bit dynamic range Prosilica GigE strobe-lighted cameras, one perpendicularly downward-looking and one forward looking (~35°). Imagery from the downward-looking camera can be analyzed to characterize the benthic communities while the forward-looking cameras are used to collect species-specific abundance and length information. Combined, these 2 imagery data sets can be used to create spatial species-specific abundance, biomass, and length-frequency distributions, along with the benthic communities around which they associate. An onboard Seabird model 49 FastCat CTD records temperature and salinity data along the AUV track, providing further environmental insight. Data Files: Video data were recorded on a digital video recorder. The position of the AUV was tracked and recorded and using a Tracklink 1500MA ultra-short base line underwater acoustic tracking system or USBL. File naming convention: Each video file is given a name consisting of a 6-digit designation for the date of the start of the dive, in UTM time, followed by an underscore and then another 6 digits indicating the start time in hours, minutes and seconds. Anotther underscore is followed by A0 or A1 for the left and right video cameras respectively. For example, the video file 120418_001003_A0.mpg was from a dive on April 4, 2012 that started at 10 minutes and three seconds after midnight and was recorded by the left video camera. It is not possible to have a timestamp displayed on the video that records time more precisely than to the nearest second. However, for stereo pair analysis to measure fish length, a higher temporal resolution is required to ensure both cameras are displaying a given fish at exactly the same time. For this reason, video analysts record the the time of a sighting (and measurement) of a fish based on the elapsed time from the start of the video, recorded in minutes with a precision of 4 significant digits. Resource Description: Spreadsheet of results of analysis of video imagery that is geo-referenced using time to the associated navigation file. Resource Description: Digital video and still imagery that is geo-referenced to navigation files.
ground condition
Please acknowledge the NOAA Coral Reef Ecosystem Division, Pacific Islands Fisheries Science Center as the source of this information.
NOAA IRC
NMFS/PIFSC/CRED
1845 WASP Blvd., Building 176
Pacific Island Benthic Habitat Mapping Center, Coral Reef Ecosystem Division (CRED), Pacific Islands Fisheries Science Center (PIFSC), NOAA
Unspecified
Complete
The horizontal position accuracy for the AUV position is estimated at 40 meters. The primary sources of this error is due to uncertainty in the position of the vehicle as determined by the USBL system. A second but smaller source of error was due to uncertainty associated with the Sette's position as recorded by the vessel's GPS system.
Video data were recorded on a digital video recorder. The position of the AUV was tracked and recorded and using a Tracklink 1500MA ultra-short base line underwater acoustic tracking system or USBL. Each video file is given a name consisting of a 6-digit designation for the date of the start of the dive, in UTM time, followed by an underscore and then another 6 digits indicating the start time in hours, minutes and seconds. Anotther underscore is followed by A0 or A1 for the left and right video cameras respectively. It is not possible to have a timestamp displayed on the video that records time more precisely than to the nearest second. However, for stereo pair analysis to measure fish length, a higher temporal resolution is required to ensure both cameras are displaying a given fish at exactly the same time. For this reason, video analysts record the the time of a sighting (and measurement) of a fish based on the elapsed time from the start of the video, recorded in minutes with a precision of 4 significant digits. Spreadsheet of results of analysis of video imagery that is geo-referenced using time to the associated navigation file. The video was paused when a fish was seen and comments were inputted directly into the a spreedsheet on how many fish there were.
none
These data are not to be used for navigational purposes. NOAA makes no warranty regarding these data, expressed or implied, nor does the fact of distribution constitute such a warranty. NOAA cannot assume liability for any damages caused by any errors or omissions in these data, nor as a result of the failure of these data to function on a particular system.
Fish count record is in excel. List of navigation files: Trackline data is in .csv format.
