Visual image event detection technology exploration
Project Manager/Lead Engineer: Duane Edgington
Lead Scientist: Bruce Robison
In this project we propose to investigate the means to visually process
images for event detection and for the recognition of target biological species,
with these systems to be deployed on autonomous underwater robotic vehicles (AUVs).
This is a technology exploration project, to review both state of the art and
state of commercial practice in detecting events (new features in the visual
field) using video cameras, digital still cameras (CCD pixel arrays) and
neuromorphic artificial retinas and photoreceptor arrays.
We seek to develop vision systems to add to AUVs to detect when they pass an
organism, and to take an image of the organism. Our dream is to be able to
identify the organism autonomously in real time.
We will evaluate application of feature-detection video frame grab technology
developed by Steve Rock’s group being adapted to the ROV Tiburon.
We will evaluate underwater and commercial digital still cameras, with
illumination by strobe, as a potential lower power alternative to continuous
light and video on AUVs. We will also evaluate state of the art feature and
motion detection chips modeled after biological vision systems. These
neuromorphic systems feature ultra-low power, large dynamic range and intrinsic
real-time image processing and feature detection. Our application focus will
include both midwater and deep ocean animals as well as animals that use
bioluminescence in the deep sea.