Commonly used denoising algorithms include median filtering, Gaussian filtering, mean filtering, wavelet denoising, nonlocal mean denoising and total variation denoising.

1, median filtering

Median filtering is a nonlinear digital image processing algorithm, which is used to reduce or eliminate noise in digital images. By sorting the pixel values in the moving sliding window of the image, the middle value is taken as the new value of the current pixel.

This filtering method is widely used to remove salt and pepper noise, in which the single pixel value in the image is strongly disturbed. The advantage of median filtering is that it can effectively preserve the edge features of the image, while removing the influence of noise, and will not introduce too many fuzzy effects.

2. Gaussian filtering

Gaussian filter is a linear smoothing filter, which is used to denoise and blur in image processing and signal processing. Based on Gaussian distribution function, it blurs the image by weighted average of pixels in the image.

The standard deviation of Gaussian filter determines the width of the filter. The larger the standard deviation, the wider the filter distribution and the more obvious the smoothing effect. Gaussian filtering is usually used to remove Gaussian noise, which has a good effect on some continuous noise models.

3. Mean filtering wavelet denoising.

Mean filtering wavelet denoising uses wavelet transform to decompose a signal or image into subbands with different frequencies, and then denoises these subbands through threshold processing.

Wavelet denoising decomposes the image by wavelet to get subbands with different frequencies. Threshold processing is performed on each subband to remove noise components. The processed subbands are reconstructed into denoised images by inverse wavelet transform.

4. Non-local mean denoising

Non-local mean denoising is an image denoising algorithm, which is based on the concept of similarity and reduces noise by finding similar blocks in the image. Non-local mean denoising mainly uses the information of similar blocks (areas with similar textures or structures in the whole image) and local information near pixel points.

5. Total variation denoising

Total variation denoising is an image denoising algorithm. By calculating the sum of absolute values of differences between adjacent pixels in an image, the total variation of the image is minimized to smooth the image, thereby removing noise.