September 27, 2012

A web application for the manual identification of marine species in seafloor imagery. Brown Bag Seminar with Serrano Pereira, Visiting Student, MIT Sea Grant

Date: Tuesday, November 13, 2012
Time: 12:00 - 1:00 pm
Location: MIT E38-714

Bring your lunch and learn about the Marine Species Identification System tool, a web application for the manual identification of marine species in seafloor imagery, and its application for estimating percent coverage of Didemnum vexillum.

Didemnum vexillum is a sea squirt that is non-native to the Atlantic coast of North America and has the potential to cause significant changes to the structure and biodiversity of areas where it is abundant. Its threat to groundfish and scallops in important fishing grounds like Georges Bank is also of major concern. The Odyssey IV AUV was deployed to survey areas in Georges Bank, Stellwagen Bank and Nantucket Sound from 2008 through 2010 as part of ongoing studies and images of the seafloor were collected. The image data from these surveys can be analyzed for the presence of D. vexillum in order to estimate its range and spatial coverage.

Current methods for mapping the spatial coverage are based on the human analysis of images using software that is not optimized for this task. Marine Species Identification System (MaSIS), a species identification tool, was developed to address this problem. This tool makes it easier to identify organisms on seafloor imagery and to calculate the percent coverage of species of interest. Image data from earlier surveys will be analyzed with this tool to estimate percent coverage of D. vexillum.

About the Brown Bag Lunch Series
MIT Sea Grant staff and students present projects, programs, research, and other topics in an informal setting to inform and inspire MIT colleagues and the wider community. The Brown Bag Lunch Series is a monthly event, and open to the public.


Serrano Pereira, a visiting student from The Netherlands, developed a tool to analyze imagery data efficiently.

Quick Links