Deriving quantitative metrics from OOI high-definition video data for the purpose of automated QA/QC

The Ocean Observatories Initiative (OOI) has a SubC 1Cam high-definition video camera (CAMHD) delivering real-time data since August 2015 from inside the caldera of Axial Seamount volcano, approximately 450 kilometers off the coast of Washington. These data are archived on the OOI's raw data repository at Rutgers University and are publicly available for download via the Apache HTTP Server endpoint [1][2]. Due to the high volume of video data being collected, large individual file size and limited data interaction capabilities provided by the HTTP endpoint, we leverage community developed tools and value added data products for the purpose of automated video data quality assurance, control and assessment [3]. In this paper, we illustrate methods for automated QA/QC of OOI video data, built upon the tools and data products generated under the National Science Foundation (NSF) Ocean Technology and Interdisciplinary Coordination (OTIC) sponsored program Cloud-Capable Tools for MG&G-Related Image Analysis of OOI HD Camera Video [4]. These include CamHDMotionAnalysis [5], PyCamHD [6], and the CamHD Compute Engine [7]. CamHDMotionAnalysis uses photometric video flow motion analyses techniques to identify and timestamp photometrically unique sections in each video, thereby producing a powerful metadata record that can be used for targeted scene extraction and content analysis. PyCamHD allows users to programmatically analyze, extract and compile content from each video on the archive into derived data products, such as time lapse videos, cluster counts of chemosynthetic bacteria floating in the water column (floc) or a quantification of bacterial mat coverage at the base of the vent. CamHD Compute Engine provides users with a computational platform to rapidly develop, share and execute code on a locally hosted copy of CAMHD raw video data files via a JupyterHub instance [8]. Here we illustrate how these tools and their products can be applied for the purpose of automated QA/QC on oceanographic video datasets following a routine and standardized sampling protocol, such as that produced by OOI's CAMHD.