Adaptive Median Filtering using MATLAB
MATLAB is a great tool. I recently wrote a code to perform an adaptive median filtering on a somewhat noisy image. The code works great in removing noise on an image…. It is a simple code that I wrote in a few minutes using MATLAB… for simplification the code does not render regions near the edges… the spatial mask which is a three by three window could not process the edges since it is out of bound…. I could have use padding and pad the edges so that the spatial mask can process the edges…. but you get the idea…and the code…Adaptive Filter
Adaptive median filtering is not the only way to remove noise on an image. It depends on the what kind of noise exists in the image. For example, if the image has a salt noise only, then using contraharmonic mean filter with Q=-1.5 probably will do the job, or if you have a pepper noise only, then using contraharmonic mean filter with Q=1.5 probably will do the job.
Also, extracting noise from an image, not necessarily we have to use time domain or spatial domain. We can also use the frequency domain such as by using band reject filter or notch filter to reject noises at certain region of the fourier spectrum.
49 Comments May 27, 2007