当前位置：网站首页>Why is it necessary to divide the variance by 255^2 when adding Gaussian noise using the imnoise function of MATLAB

Why is it necessary to divide the variance by 255^2 when adding Gaussian noise using the imnoise function of MATLAB

2022-04-23 03:38:00 【Blue area soldier】

because MATLAB For different types of image data , Normalize the variance to [0,1] Between , It can be calculated that ,uint8 Types of images , The range of variance is [0,255^2]（ This means that the maximum variance of this type of image can only reach 255 The square of , You can write the calculation formula of variance and have a look , Of course, no image can achieve this value , It's just that according to the calculation formula of variance and the value range of image, the maximum can reach ）, Then the variance value you want to add is 100, Namely
imnoise(I, ‘gaussian’, 0, 100/(255*255)).
And for double Types of images ,MATLAB The default value for the image is [0,1], Of course, you can be greater than 1, But more than 1 Value of imshow It's all white , Treat as a maximum . The image value is already in [0,1] Between , So you don't have to scale .

版权声明
本文为[Blue area soldier]所创，转载请带上原文链接，感谢
https://yzsam.com/2022/04/202204220602025466.html

边栏推荐

猜你喜欢

随机推荐