Separation of Images into Waviness and Roughness Images by Filtering
The example shows how an image can be separated into Waviness and Roughness images by use of a large Gaussian Filter kernel. This might often be desirable when measuring roughness within specific wavelength intervals.
Raw Image
AFM polymer image
Filter Kernel
The filter kernel has a size of 123 x 123 points and a cutoff wave length of 2000 nm when applied on the above image.
The filtering process is a convolution by the filter image and the surface image.
Gaussian Filter Kernel Image
Waviness Image
The smoothening effect of the large filter creates the Waviness image where only the long waves are seen.
Waviness image created by filter
Roughness Image
The difference between the original image and the Waviness image is the Roughness image where only the short waves are seen.

It is often desirable to measure the roughness on the roughness image rather than the raw image, which can be dominated by the long waves.
Roughness image created by Gaussian Filter
RSS 2.0