An evaluation of optical holography applied to imaging in situ plankton
This work investigates the application of optical holography as a non-destructive means to obtain accurate visual images and three-dimensional co-ordinates of in situ plankton. In-line and off-axis modes are examined theoretically within the framework of recording the subject in water and reconstructing the image in air. The focus of this work is on the design, construction and evaluation of an off-axis recording geometry that is suitably applied to variable concentrations of plankton in the 10-5 m to 10-3 m size range. Building upon previous work in underwater hologrammetry, a novel recording geometry and the limiting conditions for high resolution image play are introduced. Crucial to this approach is the reduction of field angle and aperture of the reconstructed hologram. Most of the parallax information commonly associated with a display hologram is sacrificed in favour of preserving two-dimensional image fidelity. The image volume is considered as a narrow field angle water core that can be localised at the hologram aperture within the bounds of paraxial approximation. The depth range of the recording volume is increased by the introduction of a side-lit array of subject beams that extend a considerable distance from the hologram. Image data is interrogated in two-dimensional optical sections using a computer controlled replay system and the three dimensional co-ordinates are obtained from a standardised origin. A procedure is introduced for determining the reconstructed image point co-ordinates in air for an axial subject point recorded through a water/glass/air planar boundary of arbitrary dimensions. Experimental data is obtained for the system resolution and analysed in terms of expected theoretical values following the Raleigh criterion and the aberration and speckle limiting factors of coherent imaging. Holograms recorded of variably dense plankton populations are compared to attenuation coefficient and traditional microscope counts in order to quantify system performance.