Microscopy - Point Spread Function, Focus, Resolution

The aim of this thesis was to design new algorithms for processing image data from microscopes and demonstration of the possibilities of their use on standard samples (latex particles of different diameter). Results were used for the analysis of real objects inside the living mammalian cell. For the design of these algorithms was necessary to first understand how the image in the microscope is build, including a variety of lens aberrations. It was necessary to start with simulations of ideal case displaying one point (simulation PSF). Images of Airy discs in the plane of focus, or simulations using the ENZ theory. Available ENZ simulations provide only a few sections of different focal planes. It was necessary to adjust them to a usable form for generating a full 3D view. Using these algorithms, it was examined the behavior of the basic lens aberrations, and the behavior of two particles (objects) at different distances from each other. At the conclusion of these observations, it was necessary to redefine the terms Focus and resolution. Furthermore, the definitions have been introduced for discriminability and distinguishability of objects in an image. Thanks to the new definitions and new viewing (information entropy) to challenge the discriminability/distinguishability problem of objects in the image was possible to design and develop algorithms for image processing that enable to detect objects below the Abbe resolution condition using standard optical bright field microscopy. It has been found experimentally that the limiting factor for resolution using this method is the size and resolution of the camera chip. When using a chip with a higher density of points, we can achieve better results (detection of smaller objects) using the same algorithms