Hackaday Prize Entry: Ground Penetrating Radar

This year’s Hackaday Prize is heating up, and right now there are quite a few projects in the works covering domains that are rarely, if ever, seen coming out of a garage or a workshop. One of the most interesting is [Glenn Powers]’ Open Ground Penetrating Radar. It’s exactly what the title says: an open-source radar system that can see into the Earth for less than $500.

While ground penetrating radar is great for archaeology and people searching for hoards buried in the middle of farmland, the biggest application is safety. You need only to Google “Florida sinkhole” to see the value of peering into the Earth.

[Glenn] is building his ground penetrating radar with a bare minimum of parts. A Baofeng VHF/UHF My First Radio™ serves as the signal generator, the controller is just an optoisolator, and the switch controller is a 7404 hex inverter. It literally can’t get simpler than that.

Of course these components can only be assembled into a simple radar, and the real value of a ground penetrating radar is the ability to map an area. For that, [Glenn] is bringing out a Pi and a GPS dongle to control the whole thing. Visualization is provided by none other than the US Navy. If it works for submarines, it should work for a metal cart, right?

It’s a great project, not only in the fact that it could help a whole bunch of people, but as a prime example of doing so much without tens of thousands of dollars in test equipment.

The 2015 Hackaday Prize is sponsored by:

33 thoughts on “Hackaday Prize Entry: Ground Penetrating Radar

  1. Maybe someone with more ordnance disposal experience than me could comment, but it seems like if this could fly in any fashion there are still a lot of unexploded mines in the world to be unearthed and this wouldn’t be a bad start.

      1. Why not? Better $500 dollars of cart gets blown to bits than a human bean. He does mention in the logs that it could be pushed/automated/aerial, you’d probably want one of the last two for landmine clearance. This kind of automated mapping could go a long way to clearing the world of the estimated 110 million landmines still active. 10 people a day die unnecessarily. Many charities would be interested in this project – this one has funded £1M for similar research. http://www.findabetterway.org.uk/

        1. I am definitely going to use ‘human bean’. I don’t know whether that was a typo, autocorrect or an eggcorn but it is so very cute.

          On the more serious side, how would you distinguish a landmine from a rock using this radar?

          1. I’m no expert on GPR, but I’d be willing to guess that metals show up very differently than rocks, metals are usually much denser than rocks, and would thus have a very different “signature” on the radar.

          2. @Maxwell
            A lot of landmines are made with very little metal for exactly that reason. Anti-personnel mines especially often have shells made of plastic, or even wood. There really aren’t too many options besides a) vehicles, possibly autonomous, specifically designed to set everything off, and b) a legion of steady-handed men with sticks.

            In the aftermath of WW2, German POWs were made to de-mine areas they had previously mined using technique ‘b’, and they were told in no certain terms that they’d be essentially line-dancing through the field at gunpoint afterwards. The de-mining was quite thorough.

        1. Actually I meant that in the more rhetorical sense (as in “if it actually worked”) but if you *could* get it usefully airborne that would be a good thing (and much better than some of the silly “drone”/quadcopter applications that keep popping up). In any case, yes…get the same outfits that are cranking out the ***uinos for $3 a copy to start making these and hopefully there’d be a lot fewer funerals and amputations.

          1. Perhaps a hovercraft with a soft skirt could distribute its weight evenly & gently enough to avoid setting most mines off. No idea on the range of sensitivity mines have.

    1. I’m a geophysicist who specializes in UXO. Just a pedantic point: mines and UXO are considered two different things in our field, even if they’re both technically ordnance. They’re opposite problems, actually. A mine is easy to remove but difficult to detect (by design). A UXO, which is a dropped or fired explosive that failed to detonate, is easy to detect (being typically made of metal), but difficult to safely remove. But the biggest problem with UXO is that it typically exists near lots of exploded ordnance, which is also metal. Therefore the problem is discrimination, not detection, because you cannot carefully remove every metal detection. The usual process is to use a more sophisticated metal detector that falls under the geophysical category “controlled-source EM”, or CSEM. It’s virtually the same device, though it may have multiple TX and RX antennas. Similar to GPR, but operating in the f range where induction far overpowers wave transmission and reflection. That means direct imaging is not particularly useful, so the data is inverted, and the results compared to a library of some sort.

  2. Wow are these Florida SInkholes real??? How do you manage to rebuild after that? Or the cities just let the holes there as tourists attractions? I there a count of disapeared peoples because of this phenomena?

    1. If my building were destroyed by a geology related sinkhole (not man-made) and I had enough resources to rebuild, I’d do it somewhere less likely to get a sinkhole I think.

  3. The cart isn’t metal, is it? Pretty sure every GPR cart I’ve ever seen is made of plastic. I think it’s an issue with reflections coming off the cart rather than the earth, but I’m not 100% on that.

    1. At the GPR frequency range, reflections occur on permittivity differences, not conductivity differences. Which isn’t to say that there won’t be reflections from the cart, but that it won’t matter if it’s metal or not.

    2. I see the bigger problem being the sway bar up front. The metal will give you early returns but you’re gonna get noise from the air/ground interface anyways. Having the swaybar/bridle pivot back and forth is going to make this return non-constant which makes subtracting it out complicated at best.
      The metal frame probably has a bigger impact on antenna radiance which will affect the shape of your field of view but as far as early returns are concerned as long as they’re constant you can adjust for them.

  4. Sinkholes are probably better surveyed with shallow seismic. Radar depth penetration is not very good at high frequencies. At low frequencies, the antenna gets rather large. Getting good S/N and imaging requires some serious number crunching. For low cost, the impulse Ultra Wide Band radar concept is pretty hard to beat. Making the transmitter and receiver separate simplifies construction and improves the result. Being able to change the source receiver separation (offset) is a huge improvement over coincident only. The quality of an image is dependent upon how many offsets are recorded and what the spacing is. This applies to GPR in exactly the same manner as in seismic. Typically the impulse designs will sum a large number of impulses at each location.

    Here’s an interesting design for a high power system.


    BTW GPR is now regulated by the FCC in the US. It wasn’t for a long time, but it is now.


  5. While this is awesome. GPR is relatively shallow. Especially in karst topography where the soil is fairly conductive. Most sinkhole chokes are greater than gpr penetration depth. I doubt there is going to be a lot of sinkholes found with this. Maybe UXO but most of that is found using magnetic induction surveys. What would be really awesome is capacitive coupled resistivity array with non proprietary analysis software.

  6. We’re given little more than a parts list and block diagram. No theory of operation. No raw oscilloscope traces or post-processed output by which one might infer how it works, or to show that it works at all.

    “The radio controller is a 4N25 opto-isolator.” Presumably that keys the transmit switch? I know a little about radar but not GPR, so I hit Wikipedia and found this:

    “Ground-penetrating radar uses a variety of technologies to generate the radar signal: these are impulse, stepped frequency, frequency-modulated continuous-wave (FMCW), and noise.”

    That radio’s not going to produce stepped frequency, FMCW, or noise without simulated microphone input. It will just produce a burst of CW. Which might qualify as impulse, but I suspect misusing a $30 consumer radio as an impulse generator doesn’t work very well.

  7. Kudos for the effort, but good effort and wishful thinking does not replace good engineering. There’s no way this can work with the components described. The T/R switch and RF detector (or MSO-19) are just far, far too slow to receive any echo closer than a dozen or so meters, by which time ther’e no signal left due to attenuation. Further (the real killer) the isolation (the amount of power it lets through) of that T/R switch is extremely poor — just the leakage through it will horribly swamp any echo from even the strongest reflectors. Better to pick up some properly selected components from Mini-Circuits and engineer a real solution: maybe not cheap, but not that much more expensive than the total BOM cost here.

  8. What waveform do you plan to use? Did you double-check that your TR switch can handle the TX power of your radio source? Suggest testing it in free-space first to make sure you can detect reflections of objects at expected distances, when you’ve confirmed the radar works then direct it at the ground. You should be able to get it to work.

  9. Please I want to git more information to help me how can I collect the requirements in circuit. ( I need diagram )
    In details. What’s the max. Depth in ground to detect metal as gold or copper.

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.