Research Of Confocal Microscopy Image’s Deconvolution Noise Reduction And Edge Extraction Algorithm

Confocal microscopy imaging system has the characteristics of accuratemeasurement in three dimensional space, high spatial resolution, reducing the advantagesof the focal plane light noise and so on.It is a new kind of imaging technology. But as aresult of confocal microscopy imaging system is restricted by the diffraction limit, andsome optical properties of confocal microscope group may lead to collect the image to acertain extent, deformation or distortion, the researchers observed the resolution of theimage is reduced, at the same time reduces the truth of information. To improve thequality of the confocal microscopic images and resolution, we can have a deconvolutionof confocal microscopy image denoising processing, increasing the signal-to-noise ratioof the image. Deconvolution to deal with the noise is a kind of an effective method toimprove the quality of the confocal microscopic images and resolution.This thesis analyzes the confocal microscopic image degradation mechanism andprocess, and studies four kinds of deconvolution algorithm for confocal microscopyimages: constrained least square deconvolution denoising algorithm, Lucy-Richardsondeconvolution denoising algorithm, blind deconvolution algorithm and iterative blinddeconvolution algorithm. After further study of the operation principle of each algorithm,writes the above four types of application of the algorithm and shows each algorithm isgood at field,and improves a kind of image restoration effect evaluation method.Inbiological research, the microstructure of microscopic image contour extractionalgorithm to study the Laplacian of gaussian (Gauss) operator, Sobel operator, Robertsoperator, Canny operator, and verify the effectiveness of each operator by experimentalmethods.And this thesis studies the microstructure contour extraction is an effectivequality evaluation method.This dissertation also designes and write the biological fluorescence scanningconfocal microscope image processing part of the software of singlechip processorsystem, thus improve the deconvolution noise reduction software biological confocalmicroscope. Parts of the image processing module include: image display module, imagebinarization processing module, histogram equalization deconvolution processingmodule, processing module, image microstructure contour extraction module, the RGBcolor image processing module.