Denoising Radar Image from White Noise

How Can I remove the white noise from the radar image below using a Treshold only?

Answer 1

[Raunak Gupta]

Hi,

Since you are plotting the “Doppler Frequency” vs “azimuth” on figure it cannot be directly imported as an Image Format file in MATLAB. So, you may try to save the data in an Image format or can apply appropriate filtering for white noise on “Doppler Frequency” before plotting it onto figure.

For Both Image and “Doppler Frequency” vector you may use imgaussfilt which is effective at removing white noise from image. Also I can see different coloring in the figure So, I can assume there are some intensities attached with “Doppler Frequency”. I would suggests you to save the data in Image format and then apply imgaussfilt with sigma as your thresholding parameter. You may see the ‘FilterDomain’ parameter in imgaussfilt for getting appropriate filtering.

[Rafi Ahmed]

Hello,

Thanks for your answer. I will definitely use the function. Moreover, the matrix before the image format contains the absolute power value (e.g. 0.00325675) of the signals. There is a formula to find the threshold voltage but I could not understand how to apply it on that matrix. Should I do the squared value of the voltage to get that power? If I do that, the value is significantly larger than the power values. Can you give me any idea how do I use the threshold voltage to get the threshold power?

[Raunak Gupta]

Hi, Since you are mentioning that the there is white noise present in process(i.e. The signal power), I would recommend you to use gaussian filter mentioned above with sigma as a controlling parameter. Any kind of thresholding would not remove white noise as the noise spans on entire spectrum.

If data doesn't have white noise then you can calculate the threshold power by taking square of the threshold voltage with multiplication with some constant. That constant will depend upon the equations of system you are dealing with. I am assuming that data is three dimensional with each pixel in the image represents a signal power.

Hope this helps.

[Rafi Ahmed]

Thanks a lot for your replies. It helps a lot.

[darova]

Any chance of accepting the answer?