A computer algorithm to implement linear structured illumination imaging

The conventional diffraction limit defines a finite range of spatial frequencies that can be transmitted through a microscope. To reveal more information about the objects that are observed by microscope, techniques that can go beyond this limit need to be developed. Structured illumination microscopy (SIM), one such method, uses patterns of excitation light to encode otherwise unobservable information into the observed image. Although the method has been well developed, the procedure of this technique is complicated. During the procedure, after encoding the unobservable information into the observed image, the superresolution information components need to be separated, shifted, and reassembled. These procedures have never been clearly explained. In this project, a computer algorithm of the linear structured illumination microscopy technique is developed. To implement this algorithm, multiple images of an object are taken with different phases and orientations of sinusoidally patterned illumination. Superresolution information components then can be extracted from these images. The procedures of separation, shifting, and reassembly of the superresolution information components are presented, explained, and verified. A block diagram of the whole procedure of the structured illumination method is presented. The results of the conventional microscope and the structured illumination algorithm are generated and compared. When applied to test objects, the performance of the algorithm is found to be in agreement with theoretical predictions, thus verifying the theory and the implementation algorithm. The block diagram of the whole procedure of the structured illumination and the explanation of the procedures of separation, shifting, and reassembly of the superresolution information components can be taken as the instructions of how to implement this method. This project report is intended to serve as a useful reference for researchers to understand this method.