| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 18.62 MB | Adobe PDF |
Authors
Advisor(s)
Abstract(s)
The exploration of water bodies from the sea to land filled water spaces has seen a continuous increase with new technologies such as robotics. Underwater images is one of the main sensor resources used but suffer from added problems due to the environment. Multiple methods and techniques have provided a way to correct the color, clear the poor quality and enhance the features. In this thesis work, we present the work of an Image Cleaning and Enhancement Technique which is based on performing color correction on images incorporated with Dark Channel Prior (DCP) and then taking the converted images and modifying them into the Long, Medium and Short (LMS) color space, as this space is the region in which the human eye perceives colour. This work is being developed at LSA (Laboratório de Sistema Autónomos) robotics and autonomous systems laboratory. Our objective is to improve the quality of images for and taken by robots with the particular emphasis on underwater flooded mines. This thesis work describes the architecture and the developed solution. A comparative analysis with state of the art methods and of our proposed solution is presented. Results from missions taken by the robot in operational mine scenarios are presented and discussed and allowing for the solution characterization and validation.
Description
Keywords
Underwater Images Image processing Image enhancement Color correction Computer vision Underwater robots Underwater mining