Research on Image Denoising Algorithm Based on Spectrum Information

With the continuous development of science and technology, information becomes indispensable in people's daily life. Among them, the image as a carrier plays an important role. It is difficult to directly obtain image information. We often irradiate the high-intensity coherent wave emitted by the light source onto the target to obtain the light wave with information. After propagating a certain distance to form diffraction intensity, we obtain the spectrum of the target. Finally, we perform denoising based on the spectrum. Obtain image information through two-dimensional inverse discrete Fourier transform and then obtain the target image.The spectrum information measured by over-sampled images contains some noises, which can be divided into Gaussian noise, Poisson noise, impulse noise and Rayleigh noise according to their statistical characteristics. Based on the noisy spectrum in Annexes 1, 2 and 3, this paper studies two problems through denoising methods such as mean value filter, statistical ranking filter and adaptive filter. We consider the influence of noise type, filter specification and other factors on the image restoration effect, and give the image restoration probability under different oversampling rates and prior information by constructing noisy spectrum information data. The effectiveness of the algorithm has potential application value for military and medical fields.