Filters based on min max mean median standard deviation

Hello, I have hundreds of images and I want to fitler them to retain edges with minimum noise. I want to do this automatically. Like.

//Analyze image
meanStdDev(roiImage, meanImg, SDImg);
minMaxLoc(roiImage, &min, &max);

//filter based on meanImg SDImg
if(condition_on_mean_n_SD)
{
    blur(roiImage, roiImage, Size(3,3));
    filter2D(roiImage,.....);
    imshow(......);
}
else if()
{
    blur(roiImage, roiImage, Size(3,3));
    blur(roiImage, roiImage, Size(3,3));
    filter2D(roiImage,.....);
    imshow(......);
}

my question is, what should be the conditions? PS: the code above is just an idea.