Basic principle of bilateral filtering

Bilateral filtering （Bilateral filter） It's a nonlinear filtering method , The essence is based on Gaussian filtering , The purpose is to solve the edge blur caused by Gaussian filtering . Combining the spatial proximity and pixel similarity of an image , At the same time, spatial information and gray similarity are considered , To achieve the purpose of edge preserving denoising . Simple 、 Non iterative 、 Local characteristics .

The advantage of bilateral filter is that it can save edges （edge preserving）, Gaussian filtering is generally used to reduce noise , It will blur the edges more obviously , The protection effect for high-frequency details is not obvious . As the name suggests, bilateral filter has one more Gaussian variance than Gaussian filter sigma－d, It's a Gaussian filter function based on spatial distribution , So near the edge , Pixels farther away from the will not affect the pixel value on the edge too much , This ensures the preservation of pixel values near the edge . But because too much high-frequency information is saved , For high-frequency noise in color images , The bilateral filter cannot be filtered out cleanly , It can only filter the low-frequency information better .

Effect comparison

from ：
1）https://zhuanlan.zhihu.com/p/94347017
2）https://blog.csdn.net/u013066730/article/details/87859184