Abstract

Semiautomated Classification of Acoustic Imagery Using ArcGIS and ENVI
Track: Ocean, Coastal, and Marine Resources
Authors: Bryan Costa, Tim Battista

The majority of shallow-water coral reef habitats have been successfully characterized by conducting heads-up digitizing and interpretation of high resolution imagery. These resulting maps, however, are time intensive to produce, limited by the size of the minimum mapping unit, and ultimately irreproducible because they depend on the accuracy of the person that is digitizing. In order to address these difficulties, an alternative mapping technique has been developed to automate the process of visually distinct features in these images. Initial results indicate that this new mapping delineating and attributing features on the seafloor. This technique uses ArcGIS to derive a suite of morphometrics from acoustic imagery, and ENVI to identify and attribute approach is: (1) 7x more time efficient, (2) equally thematically accurate, and (3) measurably more objective the manual approach. This approach, therefore, has the potential to increase the repeatability and efficiency with which maps are produced.

Bryan Costa
NOAA
1305 East West Highway
N/SCI12, SSMC 4, 9th floor
Silver Spring, Maryland 20910
United States
Phone: 301-713-3028
Fax: 301-713-4384
E-mail: bryan.costa@noaa.gov

Tim Battista
NOAA Biogeography Branch
1305 East West Highway
N/SCI12, SSMC 4, 9th floor
Silver Spring, Maryland 20910
United States
Phone: 301-713-3028
E-mail: tim.battista@noaa.gov