CODEWITHSUNDEEP
pythonimage-processingmachine-learningpython-imaging-library
Yuval Adam
Yuval Adam

Reputation: 165232

Image noise reduction techniques

Consider an image that looks like so:

enter image description here This image represents the output of a rain cloud radar with heavy rainfall from top to bottom, and significant amount of noise in the center area (all the specks in area from the center going outward between 2 and 5 o'clock)

What image processing techniques can be used to remove noise in this image? Relevant code will be implemented in Python/Pillow, but language agnostic answers will be accepted as well.

Upvotes: 3

Views: 1798

Answers (1)

lejlot
lejlot

Reputation: 66795

It looks like you can simply compute the connected components (CC) through basic BFS on the pixel graph, and then remove components under some area threshold value (so you only leave big, connected shapes).

For example in ImageJ (Java) there is a "parcticle filter" which will do it for you.

Another option is to run erosion filters, also available in ImageJ (and any other image processing library).

Upvotes: 5

Related Questions