You can't increase horizontal resolution once you have collected the data. There are limits to this anyway depending on the antenna frequency being used. You can improve the horizontal display resolution by using: Plot Options> PointmodeScale> PixelPerSample or PixelPerTrace.

To enhance the hyperbolic features (services), try to use a subtracting average instead of background removal. This will enhance your hyperbolic features while removing horizontal features. choose a  averaging trace number approximately equal to the length of the hyperbolas as a starting point. You may need to experiment to get the optimal number for your data, depending on how aggressive you want to go. Note, there will be some loss of amplitude but less for the hyperbolic features, and hence they will become more prominent against the background. Note also, if the horizontal features features are of interest then this step will be detrimental.

Also try removing the dewow as the background removal or subtracting average will effectively do the same thing. Process order may be important you will need to experiment here. The fewer processing steps you can get away with generally the better, as each step can add its own problems.

Let me know how you go or feel free to send me a raw data file to play with.

Hi Pranjal,

Enhanced horizontal resolution and extended hyperbolas lengths are two incompatible properties of a radargram. It's like trying to enhance the in-depth resolution without compressing the signal impulse.

We usually apply migration to collapse hyperbolas into "points" that represent the targets centers. But if you can't do without the hyperbolas tails to locate utilities, you may get something by applying migration on a separate file using different parameters for each hyperbola only if they are at different depths (as I see in your radargram).

In fact, you will usually apply different velocities for different depths due to strategraphic variation. In one file, you focus migration on a first hyperbola to collapse it while the others' tails are kept long enough, than you focus on another in a separate file.

If the strategraphy doesn't cause enough velocity variation (which is very rare), you can try to slightly change the time offset parameter from its true value to get different velocities for different depths.

This can't work with hyperbolas at close depths.

Good luck!

Thank You all for your valuable answer.

Val Sir,

find three imported data to play acquired by using GSSI Dual Frequency Antenna (300Mhz & 800 Mhz)Lines were crossing the road. there was potential drainage line at around 7m and lampost cable line between 27-28m on the road curb.

looking forward to your response

DZT files missing.

Sir,

please bring the cursor to the small square boxes. these are the DZT files. I don't why it becomes like that after sending

Really, I enjoyed this fruitful discussion. Let us look at the given data and then exchange idea.

Sure Mr. El-Behiry,

Mr. Pranjal will need to send me by email the files: [email protected]. The small squares are not working.

Will see what I can do,

Thanks,

Bilal Sir,

files are unable to send. It is showing attachment error. I think there is issue with format. I'll try again.

Dear All,

There is one more interesting thing I encountered on attached data below. for example, file 'L8_0-20', there is response of drainage channel at 2.6 meter distance, depth around 0.5 meter which continued in other parallel lines as well but the other reflections (at 4.8m, 6.4m, 9.2m, 9.9m and 12.3m) look from very near to the surface. Please suggest what could be the possible features related to such kind of reflections. Antenna frequency was 400MHz acquired on Asphalt road, there was nothing visible on surface, it was smooth flat and clean. Drainage metal cover encountered at 0.5m.

If I didn't know better I would say they are tram tracks that have been asphalted over. However, the spacing and locations are not consistent. Unlikely may be soil nails from an adjacent retaining wall. Likely, to be metal because of the long tails. Looks like they run perpendicular, but you can check this by using the velocity adaption tool. Sorry but have no other ideas.

Thank you Sir. 

Agree with Val Donazzolo, it seems to be some metallic object, may be a overhead lamp post, located across your traverse.

When I carry out data collection, I normally put a 360 camera above the centreline of the antenna, meanwhile the 360 camera also capture the image of the readout. Then, I can review the environment and verify possible interference from adjacent objects.

I had a similar situation. I guess you should use the manual gain (y) instead of the Energy decay, and the increase in the gain value should be start at the target depth.

There is an example of using the manual gain (y) in Goodman and Piro (2013).

Goodman, D. and Piro, S. (2013) GPR Remote Sensing in Archaeology. Springer-Verlag, New York.

Regards

Sorry for slow response. I have tried to process File DS144_005 800 MHz and found it wasn't too bad too start with. I've tried a couple of different sequences which gave similar results. You can also tweak the processing to enhance particular features you may be interested in such as the reinforced concrete slabs at the surface or the services below the pavement. I have attached a before and after image. The sequence I used was first to run a subtracting average over 350 traces (14m) to remove the dewow and noise bands. You could do this with a butterworth filter but then you will still need to run the subtracting average to remove the noise bands. Note that background removal in this case will introduce linear artifacts because the layers are not continuous. I then used a gain function starting at 3ns with a 5 x linear gain and a -5dB/m exponent. You could also use an energy decay gain but this applies a different gain curve on each trace and therefore you can lose the amplitude variation across the section which is sometimes important. Hope this helps.

https://www.researchgate.net/converse/api/topics/message/download/5e5f177fc7d8ab6b9b3b37cchttps://www.researchgate.net/converse/api/topics/message/download/5e5f1789a7cbaf01c272720f