Coregistered hyperspectral and stereo image seafloor mapping from an autonomous underwater vehicle
Journal Publication ResearchOnline@JCUAbstract
We present a new method for in situ high-resolution hyperspectral mapping of the seafloor utilizing a spectrometer colocated and coregistered with a high-resolution color stereo camera system onboard an autonomous underwater vehicle (AUV). Hyperspectral imagery data have been used extensively for mapping and distinguishing marine seafloor habitats and organisms from above-water platforms (such as satellites and aircraft), but at low spatial resolutions and at shallow water depths (<10 m). The use of hyperspectral sensing from in-water platforms (such as AUVs) has the potential to provide valuable habitat data in deeper waters and with high spatial resolution. Challenges faced by in-water hyperspectral imaging include difficulties in correcting for water column effects and the spatial registration of point/line-scan hyperspectral sensor measurements. The methods developed in this paper overcome these issues through coregistration with a high spatial resolution, stereo color camera, and precise modeling and compensation of the water column properties that attenuate hyperspectral signals. We integrated two spectrometers into our SeaBED class AUV, and one on-board a support surface vessel to measure and estimate the effects of light passing through the water column. Spatial calibration of the spectrometers/stereo cameras and the synchronized acquisition of both sensors allowed for spatial registration of the resulting hyperspectral reflectance profiles. We demonstrate resulting hyperspectral imagery maps with a spatial resolution of 30 cm over large areas of the seafloor that are not adversely effected by above-water conditions (such as cloud cover) that would typically prevent the use of remote-sensing methods. Results are presented from an AUV mapping survey of a coral reef ecosystem over Pakhoi Bank on the Great Barrier Reef, Queensland, Australia, demonstrating the ability to reconstruct hyperspectral reflectance profiles for a diverse range of abiotic and biotic coverage types including sand, corals, seagrass, and algae. Profiles are then used to automatically classify different coverage types with a 10-fold cross validation accuracy of 91.99% using a linear support vector machine (SVM).
Journal
Journal of Field Robotics
Publication Name
Journal of Field Robotics
Volume
35
ISBN/ISSN
1556-4967
Edition
N/A
Issue
3
Pages Count
18
Location
N/A
Publisher
Wiley
Publisher Url
N/A
Publisher Location
N/A
Publish Date
N/A
Url
N/A
Date
N/A
EISSN
N/A
DOI
10.1002/rob.21713