Once you start talking Radar / NMI and lots of high end stuff you have massive power needs for even a small system and then the materials needed are often not safe.
I have and have used a pediatric ultra sound machine to test gold bars to see if they are pure or fake
very interesting but never going to work in the field for many reasons.
Ground radar - big and heavy and lots of software let alone the source and the management of the field and exclusion zone and the operator. OK for seeing big ground disturbances and or changes in the ground. but as to separating and detecting small items - NOPE.
All comes back to physics and functionality. And physics says that the simplest way is to induce an eddy current into a metal object and to then read/detect that eddy current. To induce and eddy current is simple
the reading of that eddy current and to determine what material it might be gets harder. Hence the VLF and different frequencies such that lab testing has some level of guessing the target make up. Remembering there are 1001 variables.
As computing power increases
it does not always mean there is a similar rate of rise ability to detect the eddy currents - as eddy currents are subject to the transmit and frequencies AND the RFI that is around AND the receiving coil. And RFI is a killer and just because you can not see or hear or feel it, it is there and induces its own eddy currents in the detector and targets in the ground. You sit quietly in the bush but there is a ton of electromagnetic waves hitting you at all times. We will not even talk about the 1001 variables with the ground LOL.
The increase in CPU speed and better electronic components just allow for somewhat better processing and guessing of the target as the cpu and software can number crunch the input signals and give a tone or TID etc and all that happens as you swing and a target is exposed to the detecting field for a whisker of a second and the receive coil gets the signal to the cpu and then to the speaker and or screen - And your still swinging and moving the coil.
People forget just how much is going on in any detectors electronics to do all this, and such that your perception is that it is instantionus.
Hence why speed of swing and consistency of movement can be critical.
With machines that are focused on gold... One can drop some of the processing power as who needs TIDS
raw iron will behave differently to gold so some processing power can be used to provide a level of discrimination. But the key is the sensitivity of the receiving ccts to process the eddi currents and the CPU power that is used to clean up what could be interference (Ground noise - RFI = Threshold) and then to provide an output THAT the user may be able to interpret.
Thus thresholds and ones hearing profile is often the key for finding those bits of fly ****.
And why, where in the field that bit of fly **** may have been AND people forget about the importance of knowing the field shape. Hence why may will say an elliptical etc may find more that a round of the same size - but it is really one of the field shape and how the coil was moved over the ground.
Opps end of basic physics rant...
OH and ease of use and power requirements are often the biggest limiting factor as the manufacturers have to cater for the Majority of users and not the hard core few, if they are to sell enough to recover research costs.
PS. The above comments are subject to my emotions and own bias AND is post field experience and 30 plus years in the electronics and bio medical fields but are just my thoughts.
Yes, twas a mad night and I need to go and take some time out meds LOL... 6 months plus with out a swing in the field and Easter can not get here fast enough.