What is a Hilbert Transform in GPR? | Ground Penetrating Radar

What is a hilbert transform for GPR data?
This process creates an envelope of the GPR trace where the positive and negative polarity amplitudes responses get transformed into the total amplitude for each reflection response. This creates absolute values for each response so that the total amplitudes are comparable. This can sometimes be difficult to do when responses contain both positive and negative polarities.
In a recent development in GPR-Slice software, the Hilbert Transform can pic polarities based on the maximum amplitude band and convert each response into an envelope of the total amplitude, but keeps the dominant sign (+/-). This is a revolutionary adjustment that helps practitioners more easily interpret wave behaviors and helps reduce the range of possibilities for a given response.
If you have other data processing steps that you would like to learn about, leave them in the comments below.
If you use GPR and didn't know how this worked, then you should get further training. Check out our upcoming webinars, classes, and live workshops at LearnGPR.com.
****** SUBSCRIBE TO THE CHANNEL******
Subscribe to our channel: / learngpr
****** WHAT IS LEARNGPR******
LearnGPR.com is an online training platform for Ground Penetrating Radar that helps civil engineers, utility locators, and others learn about GPR through fun and engaging courses and coaching.
****** FREE TRAINING FROM LEARNGPR******
Go to learngpr.com to sign up for our free Introductory Training Video on Ground Penetrating Radar (GPR).
****** CONNECT WITH LEARNGPR******
Want to learn more about GPR? Follow us at:
/ learngpr
/ learngpr
/ learngpr (tag your GPR questions with #learngpr)
/ learngpr (tag us on your pics to get featured)
/ learngpr
/ learngpr

[@sudhap5131]

Hi Dan, Grateful for your videos on GPR,Very informative & closely following to learn more on GPR,Pls keep updating these videos

[@Learngpr]

Thanks so much! Appreciate you watching

[@hasanshirazi9535]

Great explanation. I am new to GPR and am watching and learning from your videos a lot. Thanks.

[@Learngpr]

Hasan Shirazi thanks so much. I hope you keep watching and find them helpful. What are you using gpr for?

[@hasanshirazi9535]

I am mainly into NDT of RCC structures. In addition to GPR I also use PUNDIT and Eddy current rebar detector.

[@haroonshaikh9771]

Very informative video. Hilbert transform based on magnitude in Radan 7 gave me good results to understand the energy loss in the subsurface due to change in lithology.

[@Learngpr]

Haroon Shaikh Thanks for the comment. That's awesome. Looking at rates of loss due to lithology is a great use of this process

[@davidyan106]

i want to know which software the Radan 7 is，can you give me a website？

[@GeelongCableLocations]

.......hmmm, ok, I'm normally pretty good with your videos, but some of that went straight over my head. Yes, it might be 2am in the morning here, but I think I better mark this video down as one that I need to watch again so it can sink in a bit more 😀

[@Learngpr]

Geelong Cable Locations do you have radan?

[@GeelongCableLocations]

LearnGPR yep, although admittedly we don't use it that much anymore. For the work we do these days we don't really need to post process data. Although, in saying that, if you do ever create an online course for radan then let me know as I will buy that off you.

[@bdrogin]

Thanks for bringing this to my attention, Daniel, I did a little digging, it looks like the Hilbert Transform is useful in dramatically isolating high amplitude correlated reflections from low amplitude uncorrelated background noise. I only bring it up because you don't mention that in the video. I bring up the correlation because my fading memory had assumed it might be some kind of matched-filter process, but it isn't. Naturally, not trying to get a video on "What is correlation?", that's for a math or signal processing course, not for GPR folk... :)

[@Learngpr]

bdrogin thanks for coming in. that's exactly right and somewhat what I meant when I said relative frequencies. You can definitely isolate potential targets from low amplitude signal of non-interest (think that's a word). It is very useful for concrete scanning like you do up north. Really can clean up images.

[@jrglidden]

Excellent and informative

[@Learngpr]

Thanks Joe! I love that you are up early training. Best way to start the day!

[@jinglunfeng9624]

Hi Dan, great video! I learned a lot from you. I have a question about the A-scan formation, why there are always 3 peaks/bands generated when gpr signal enters into the new material? And what does the positive and negative peak stands for? Thanks!

[@Learngpr]

Thanks fornthe comment and the question. Since it's a pulsed wave, it will have peaks and valleys that approximate a sin curve. The sequence of peaks (+) and valleys (-) indicate whether the wave transitioned into a material with a higher or lower dielectric permittivity. So analyzing the polarity can give some indication about what material the wave traveled into. Hope this helps and thanks again!

[@jinglunfeng9624]

@@Learngpr Hi prof, thanks for the quick response! I understood the wave simulates the sin wave, but why there are three rather than just two peaks generated every time the signal goes into a different dielectric material. Thanks again!

[@prospekteur8515]

Hello Dan, fantastic explaination;-) Watched this video several times in the past;-)One Question: So the Hilbert Transform is the same as the Envelope-filter ?

[@Learngpr]

Prospekteur yes. I think different software packages may call them different things. But they are the same. Thanks for watching!!

[@prospekteur8515]

Thanks, i work with ReflexW and there is Hilbert-Transform called Envelope-filter. Some other functions also named a bit different😉. Would be nice if you would explain " Dc-shift " and "deconvolution" in a Video in the future.

[@Learngpr]

My pleasure! I'll see what I can do.

[@randomstuffs7757]

please make videos explaining the key concepts of Fourier , Laplace and Z-transform, and also convolution, deconvolution, correlation and autocorrelation. thanks

[@Learngpr]

Praveen Singh thanks for the suggestions. Most of those are high level. I'll hopefully get to them at some point. We just did one about deconvolution that you should check out.

[@kemalorman9162]

How do we understand the underground space? Do we understand such gaps by looking at the signal amplitude? eg rock hole? or the burial space under the ground? thanks

[@mohammedalobaidi4109]

thank you man great ;do you have any information about wevlet transform?

[@Learngpr]

Mohammed Alobaidi thanks for the question and the compliment. Can you be more specific?

[@mohammedalobaidi4109]

LearnGPR any information About signal processing wevelet because my master I choose GPR signal processing and this is my mail if you can help me 😊 [e-mail]

[@Josh-ch4df]

3:37 - one for the blooper reel haha. Awesome video, very informative! Hopefully GSSI takes a leaf out of Goodman's book and implements polarity adjusted Hilbert Xform.

[@Learngpr]

That would be awesome. Radan has a nice hilbert transform feature. Would be great if you could just press a button and have it polarity specific.

[@pankajprasad5339]

how to explain phase and frequency in hilbert transform?

[@ouabelouabel7433]

Well very good speech but let me let you know that hilbert transform can give you apparent polarity, instantaneous frequency, instantaneous phase etc...

[@Learngpr]

Thanks for letting the rest of the viewers know. You are totally correct as there are a number of things that hilbert transform can do, but almost ALL of the users that I work with (i would estimate 99.5 percent) would ONLY be interested in the magnitude function, if they used this step at all. The point of the video was to introduce people that never used or heard of hilbert transform to understand its helpfulness in comparing relative amplitudes. Thanks again for the comment. Much appreciated.

[@ouabelouabel7433]

LearnGPR many thanks. I will post some key notes about what we can do to perform more the interpretation by using some others attributes. Thank you for your reply and for your efforts