Pairing side scan sonar imagery and ground-truth videos to characterize western Lake Michigan

Session: 60. - Seeing Below the Surface: Quantifying the Underwater Environment with Image Analysis

Charles Menza, NOAA, [email protected]
Will Sautter, NOAA / CSS, [email protected]
Ayman Mabrouk, NOAA / CSS, [email protected]
Sarah Hile, NOAA / CSS, [email protected]
Matt Kendall, NOAA, [email protected]

Abstract

Remote sensors, such as side scan sonars and LiDAR (Light Detection and Ranging) systems, can efficiently characterize the lakebed over broad areas and at fine spatial resolutions. These maps are commonly paired with in situ data to ground-truth the remotely sensed image patterns. This talk will discuss an approach to efficiently collect in situ information and reconcile the distinct spatial scales of ground-truth data and remote sensing imagery. We collected over 350 videos and 60 square kilometers of side scan data to provision the proposed Wisconsin-Lake Michigan National Marine Sanctuary with a lakebed map, and we found that ground-truth site positioning, map scale, remote sensor resolution, benthic classification types and benthic heterogeneity must all be reconciled to accurately map benthic features.

1. Keyword
Lake Michigan

2. Keyword
Spatial analysis

4. Additional Keyword
lakebed map