For marine sciences occasionally there’s a need to carry out under water photography. Optical properties of light result in serious quality problems for underwater photography. Light of different energies is absorbed at very different rates under water leading to important bluishness of the images. If the colour dependent attenuation under water could be correctly calculated it should be possible to use computerised image processing to colour appropriate digital images using Beer’s Law. In this dissertation we’ve created such estimation and correction methods that have become a lot more complex and more accurate giving successively superior correction final results. A process of estimation of downwelling attenuation coefficients from multi or hyper spectral data is a basis for automatic colour restoration of underwater taken images. The outcomes reveal that for each diving site the unique and precise coefficients can be acquired. All standard digital camera models have built-in white balancing and colour enhancement functions built to make the pictures as eye appealing as possible. These functions can in most cameras not be switched off and the algorithms used are proprietary and undocumented. However, these types of enhancement functions can be approximated. Applying their reverse creates un-enhanced images and we demonstrate that our algorithms for underwater colour correction works considerably better when applied to this kind of photos…
Contents: Colour Correction of Underwater Images Using Spectral Data
1 Introduction and Objectives
1.1 Goal and focus
1.2 About this thesis
2 Background
2.1 Light field
2.2 Optical properties of water
2.2.1 Inherent vs apparent optical properties
2.2.2 Diffuse attenuation coefficient
2.2.3 Dissolved organic matters and suspended inorganic matters
2.3 Underwater data collection
2.3.1 Spectralon
3 Colour image processing for underwater applications
3.1 Colour perception
3.2 R, G and B in digital cameras
3.3 Camera’s enhancement functions
3.4 Colour correction
3.4.1 Underwater image restoration
3.5 Implementation
4 New methods and applications
4.1 Paper I, II
Estimation of diffuse attenuation coefficients for three spectral channels in a digital camera
4.1.1 Methods and experiments
4.1.2 Results
4.1.3 Conclusions
4.2 Paper III, IV
Estimation of camera enhancement functions and preprocess-ing of images
4.2.1 Methods and experiments
4.2.2 Results
4.2.3 Conclusions
4.3 Paper V
Dissolved particles influence on Kd values
4.3.1 Methods and experiments
4.3.2 Results
4.3.3 Conclusions
4.4 Paper VI, VII
Spectral data applied on colour underwater image correction
4.4.1 Methods and experiments…
Source: Uppsala University Library
Download URL 2: Visit Now