Traitement et analyse quantitative d'hologrammes numériques

Abstract : A hologram of micro-objects contains information about the size, shape and 3D localization of the objects. This information is coded in the hologram by the frequency and amplitude modulations of the interference fringes. The extraction of this information is classically done by analyzing the 3D volume reconstructed with a Fresnel transform, a fractional Fourier transform or a wavelet transform. However, several types of artifacts appear in the reconstructed images: distortions close to the image boundaries, multiple focusing, twin-images.

The analysis of the issue of digital hologram processing under the light of signal processing and deconvolution makes it possible to give a unified description of previous contributions found in literature. We suggest an ``inverse problem'' approach based on the physical model of hologram formation to circumvent the artifacts of reconstructed images. The proposed algorithm detects micro-particles in a volume 4 times larger and with a precision improved by a factor 5 compared with classical techniques.

On-line processing of digital holograms requires fast processing algorithms. We have developed a stereological approach to estimate the mean size or the distribution of directions of a population of micro-objects.
Complete list of metadatas

https://tel.archives-ouvertes.fr/tel-00282661
Contributor : Loïc Denis <>
Submitted on : Wednesday, May 28, 2008 - 9:35:51 AM
Last modification on : Monday, January 13, 2020 - 5:46:02 PM
Long-term archiving on: Friday, May 28, 2010 - 5:51:56 PM

Identifiers

  • HAL Id : tel-00282661, version 1

Citation

Loïc Denis. Traitement et analyse quantitative d'hologrammes numériques. Interface homme-machine [cs.HC]. Université Jean Monnet - Saint-Etienne, 2006. Français. ⟨tel-00282661⟩

Share

Metrics

Record views

689

Files downloads

4111