Collecting fish counts and lengths from stereo-video systems

by William Misa, Audrey Rollo, and Jeff Anderson

At the conclusion of a fishery-independent sampling methods cruise such as SE1402, videos collected from stereo-video systems (BotCam; AUV) are processed by optical analysts in the Science Operations Team. Fish identification, counts, and lengths are generated during the video annotation process for use by the Stock Assessment Program in sampling gear comparison studies and fisheries monitoring. By recording fish in their habitat, video sampling offers a non-extractive alternative to traditional fisheries surveys.

Stereo-video platforms (BotCam and AUV)

The BotCam (Picture 1) is a stationary baited remote underwater stereo-video system designed by D. Merritt and the Pacific Islands Fisheries Science Center Coral Reef Ecosystem Division (PIFSC-CRED). The system was designed to capture bottomfish species in their habitats across a variety of substrate types and slopes at depths of 100–300 meters. The BotCam records videos onto a digital video recorder using a pair of ultra-low light cameras. Further details on this system can be found at http://www.pifsc.noaa.gov/cred/botcam.php.

Picture 1 –  Bottom camera bait station (BotCam)

Picture 1 – Bottom camera bait station (BotCam)

Please refer to the blog post entitled “SE-14-02 AUV: MHI Comparison of Fishery-Independent Sampling Methods Project” for more information about how autonomous underwater vehicles (AUVs) are used to collect data for these analyses.

Counting and measuring fish in EventMeasure

All fish species seen on video are identified to the most specific taxonomic level possible. For each species, relative abundance (MaxN) and lengths are recorded in the EventMeasureTM software package (SeaGIS Pty. Ltd., Victoria, Australia). MaxN is an estimator of fish abundance generated using the single highest count of a given fish species seen at a single point in a video recording. The optical analysts review each video set to determine MaxN for a given fish and take measurements when possible. Depending on the stereo-video platform used, habitat data, observations on marine flora and fauna, and other information derived from the video are also recorded.

Picture 2 – Counting fish in EventMeasure

Picture 2 – Counting fish in EventMeasure

Picture 3 – Measuring fish in EventMeasure

Picture 3 – Measuring fish in EventMeasure

Target fish species

While stereo-video processing involves generating counts and lengths for all fish species seen on video, surveys in the Hawaiian Islands mainly target the Deep 7 bottomfish grouping and other local bottomfish species. The Deep 7 are a commercially valuable group of bottomfish which includes the crimson jobfish (opakapaka, Pristipomoides filamentosus), lavender jobfish (kalekale, Pristipomoides sieboldii), oblique-banded snapper (gindai, Pristipomoides zonatus), deep-water red snapper (ehu, Etelis carbunculus), deep-water long-tail red snapper (onaga, Etelis coruscans), rusty jobfish (lehi, Aphareus rutilans), and Hawaiian grouper (hapuupuu, Hyporthodus quernus). Other bottomfish species of interest are the green jobfish (uku, Aprion virescens), Randall’s snapper (Randallichthys filamentosus), greater amberjack (kahala, Seriola dumerili), and almaco jack (kahala, Seriola rivoliana). Because the BotCam and AUV stereo-video platforms make use of ultra-low light cameras, color is absent from video imagery and optical analysts have to make use of body shape, gray-scale coloration patterns, and other unique characteristics to identify fish (see Video 1).

Use of stereo-video data

In Hawaiʻi, baited video camera systems have been used to identify and survey juvenile deepwater snapper habitats, as an alternative method to fishing to compare fish relative abundance to Catch Per Unit Effort (CPUE), in studies of deepwater bottomfish and their habitats, and to assess the effectiveness of deepwater marine protected areas. Current projects involving stereo-video systems in the Pacific Islands Fisheries Science Center look into the efficiency of fishery-independent sampling techniques and the operational use of fishery-independent methods in bottomfish surveys.

This entry was posted in coral reef ecosystem, Fisheries Research and Monitoring, Scientific Operations and tagged , , , , , , , , . Bookmark the permalink.