An improved non-local means filter for image mixed noise
Shuang Zhang, Lu Wang · 2021 China Automation Congress (CAC) · 2021
The non-local means filter (NLMF) can improve the denoising effect of the image while retaining good edge details. However, the smoothing coefficient h needs to be manually determined before processing. In advance, the non-local means filter has a weak denoising ability for impulse noise and SPN salt-and-pepper noise. In order to overcome these problems, this letter improves the non-local means filter for image noise, with combining with median filtering, and using an adaptive smoothing coefficient h. The proposed method can be used to mixed noise. Gaussian noise, impulse noise, or salt-and- pepper noise are processed in this letter. Compared with median filter, mean filter and standard non-local mean algorithm, the adaptive non-local means filter proves the effectiveness.