Recently, stimulated emission depletion microscopy has achieved high resolution in fluorescent imaging. In this paper, we present effects of a pupil filter on the performances of stimulated emission depletion microscopy. In stimulated emission depletion microscopy, a saturated zero-centered spot is usually used to achieve a high lateral resolution. Using a half-coated phase plate, a zero-centered spot was made with a narrow and steep gap at the center. Numerical and experimental results show that by simply inserting a central obstacle as a pupil filter, it is possible to reduce the central gap of the zero-centered spot. However in order to compensate inevitable loss of light, which is blocked by the obstacle, increased laser power is required.
Confocal Scanning Microscopy (CSM) is very useful to reconstruct 3D image of Bio-cells and the objects that have specification shape in higher axial and lateral resolution and widely used as measurement instrument. A 3D reconstruction is used to visualize confocal images and consists of following processes. The First process is to get 3D data by collecting a series of images at regular focus intervals (Optical Sectioning). The Second process is to fit a curve to a series of 3D data points each pixel. The Third process is to search height information that has maximum value from curve-fitting. However, because of various systematic errors (NOISE) occurred when collecting the information of images through Optical Sectioning and large peak deviation occurred from curve-fitting error, high quality 3D reconstruction is not expected. Also, it takes much time to 3d Reconstruction by using many 3D data in order to acquire high quality and much cost to improve signal-to-noise (SNR) using a higher power laser. So, we are going to define SNR, peak deviation and the order of curve-fitting as important factors and simulate the relation between the factors in order to find a optimum condition for high quality 3D reconstruction in Confoal Scanning Microscopy. If we use optimum condition obtained by this simulation, using a suitable SNR and the suitable number of data and the suitable n-th order curve-fitting, small peak deviation is expected and then, 3D reconstruction of little better quality is expected. Also, it is expected to save.