Ground Change Detection - specs

[Aware]

Hi All,

I have discussed this before in another thread elsewhere. However as its often UAV specific .. you may be more of a targeted reader.

I am looking for information on GCD, specifically US DoD specs, ConOps, Display requirements etc. Currently It seems that there are not a lot of products on the market. And since the media suggests there is a strong demand, either there are secret in-house DoD developments or no one has been able to crack this area?

The GCD that I am envisioning is that created from UAVs with successive fly-overs of such things as ally compounds or main roads in places such as Afghanistan. But currently this appears to be the level of information available for DoD requirements. Does anyone know if there is a white-paper, a request for tender or any specific document detailing specific current requirements for GCD? Granted it does not need to be done from a UAV, however dull, dirty boring tasks seem to come under this category. GCD from UAVs would really benefit from stable pics so maybe this is more limited to tier 1 and 2 UAVs?

Any thoughts people?

p.s. I am looking at Ground Change Detection requirements (technology independent). That is, most GCD, or GMTI, or Coherent Change Detection systems imply the use of SAR radar. I am involved with a development using EO MTI on UAVs. That is analysing plain Electro Optical sensor imagery rather than radar. And so I am interested in finding any open ended spec that doesn't rely on analysis being performed by SAR imagery.

Thanks in advance.

 

[weasel1962]

Re:

Hi All,

I have discussed this before in another thread elsewhere. However as its often UAV specific .. you may be more of a targeted reader.

I am looking for information on GCD, specifically US DoD specs, ConOps, Display requirements etc. Currently It seems that there are not a lot of products on the market. And since the media suggests there is a strong demand, either there are secret in-house DoD developments or no one has been able to crack this area?

The GCD that I am envisioning is that created from UAVs with successive fly-overs of such things as ally compounds or main roads in places such as Afghanistan. But currently this appears to be the level of information available for DoD requirements. Does anyone know if there is a white-paper, a request for tender or any specific document detailing specific current requirements for GCD? Granted it does not need to be done from a UAV, however dull, dirty boring tasks seem to come under this category. GCD from UAVs would really benefit from stable pics so maybe this is more limited to tier 1 and 2 UAVs?

Any thoughts people?

p.s. I am looking at Ground Change Detection requirements (technology independent). That is, most GCD, or GMTI, or Coherent Change Detection systems imply the use of SAR radar. I am involved with a development using EO MTI on UAVs. That is analysing plain Electro Optical sensor imagery rather than radar. And so I am interested in finding any open ended spec that doesn't rely on analysis being performed by SAR imagery.

Thanks in advance.

The question is a strange one. DoD requirements are generally constraint based eg range, type of target, type of background, angle of incidence, maintenance, useability, resolution etc. To achieve that may require differing source requirement eg wavelength type (eg polarisation, phase, frequency etc), revisit rates, 3-D/2-D (all of which don't really interest the end-user).

You can see that in something like the below.
http://www.darpa.mil/IPTO/solicit/baa/BAA-10-07_Mod01.pdf

You can search for more examples as the darpa.mil website publishes all its solicitations.

Example of an gmti brochure.
http://www.iai.co.il/sip_storage/files/6/34036.pdf

There are already a lot of research papers on GMTI, sar/inverse sar techniques etc which can be accessible via any university search function in a university library.

Now what radar does is to map using radar waves the curvature of the target area by calculating the time it takes for a radar beam to beam back from the emitter to the receiver. With multiple beams, one can identify differences in the curvature that using a computer algorithm identifies whether the object is moving or mere background noise eg surface waves. I would guess EO techniques would probably involves comparison of differences in static pictures which principally means a difference in frequency (between visual and radar wavelengths).

I would guess orgs like DARPA are more interested in what can be achieved first before talking about operational parameters eg mtbf, housing, display etc.

 

[Aware]

The question is a strange one. DoD requirements are generally constraint based eg range, type of target, type of background, angle of incidence, maintenance, useability, resolution etc. To achieve that may require differing source requirement eg wavelength type (eg polarisation, phase, frequency etc), revisit rates, 3-D/2-D (all of which don't really interest the end-user).

.... There are already a lot of research papers on GMTI, sar/inverse sar techniques etc which can be accessible via any university search function in a university library.

.... I would guess orgs like DARPA are more interested in what can be achieved first before talking about operational parameters eg mtbf, housing, display etc.

Thank you for a great and specific answer!

I downloaded your URL examples, and thanks for the Darpa search suggestion. The question is not so strange when you consider that the company that I am researching for uses EO based solutions, and most university papers, and even your darpa link are all SAR mathematical intense documents. With very sparse mention as to how the systems 'should' be employed.

I have an electronics background so I appreciate the concepts you raised, however they are similarly based on SAR, not solving a problem per se. And this problem is general GCD - not specifically combating IEDs. The application I am musing over is general sweeps of areas hours or days apart to help identify new or missing people/ assets on over-laid images. But I would really prefer knowing a more specific Conops, or resolution required etc - if it exists, for general GCD (not SAR reliant).

I understand that all the research is directed at SAR because of its cloud and foliage penetrating abilities, but of course with SAR comes complex hardware, maintenance, complex processing and cost. My clients alpha version based on EO GCD does not penetrate cloud, but at an order of magnitude or so less expense , and not involving new hardware (uses existing EO sensors) it 'should' be ideal for tier 1 and 2 UAVs which often fly below clouds anyway. I also appreciate that DoD seems to have gone to the premium end of the market because previous EO MTI/GCD systems didn't exist or their MTI were often not very sensitive or would false alarm.

I am guessing that all of this history has made DoDs efforts focus on the premium end of the market, and believing that EO MTI GCD was dead, but this is not so. SO, if there is a non sensor specific spec out there for GCD, or a guide as to the ConOps or images they require to analyse a situation, it would be better knowing that now than wait for Beta development to find that out!

I have read info in 'unmanned systems' magazine, C4ISR, even wired magazine but none of them can point to a defined GCD method that DoD requires to do the job. That either means that GCD is so new that they don't wont to constrain inventiveness, that they don't know, that its a state secret or a mix of these options.

Just thought I would put the call out there .. in the mean time I thank you for the docs you pointed me to. Perhaps that is the best that is out there...

 

[Aware]

GCD research update - IED

in addition to my last post, it would appear that 'ground change' Ops are not specifically GCD.

I reviewed the Darpa paper and their request for the SAR system was more for MTI with vectors and predicting short term tactical info such as where the target is likely to be in minutes to an hour or so. Vague reference to GCD in the past has been more about changes in assets or dismounts from an area of interest over a longer time. That is, roadside clear one day, 24 hours later vehicles detected. Since this kind of analysis is usually done post the second fly over, being able to detain suspicious dismounts is not always possible.

Upon further reading, It appears that there is no specific DoD GCD requirement, but that everyone is scrambling to provide solutions for detection of IEDs. This probably means that Detecting IEDs should be the core focus of any GCD system that is built. Again since there appears to be no specific ConOps released about IED detection, it would appear that it is up to the manufacturer to state their case of how they think it should be done, and use the sensor resolution to the best of their abilities to increase positive detections.

SAR is great when it uses ground penetrating capabilities and the ability to detect metal wires and pressure plates, then have another sensor zoom in for confirmation, however cost means that SAR is still usually limited to very main roads and expensive tier 3 platforms. It appears that there is still a niche for an effective EO IED detection solution on tier 1 and 2 platforms, even it it is as backup or confirmation as they can fly lower and more regular passes.

Any views?

 

[weasel1962]

in addition to my last post, it would appear that 'ground change' Ops are not specifically GCD.

I reviewed the Darpa paper and their request for the SAR system was more for MTI with vectors and predicting short term tactical info such as where the target is likely to be in minutes to an hour or so. Vague reference to GCD in the past has been more about changes in assets or dismounts from an area of interest over a longer time. That is, roadside clear one day, 24 hours later vehicles detected. Since this kind of analysis is usually done post the second fly over, being able to detain suspicious dismounts is not always possible.

Upon further reading, It appears that there is no specific DoD GCD requirement, but that everyone is scrambling to provide solutions for detection of IEDs. This probably means that Detecting IEDs should be the core focus of any GCD system that is built. Again since there appears to be no specific ConOps released about IED detection, it would appear that it is up to the manufacturer to state their case of how they think it should be done, and use the sensor resolution to the best of their abilities to increase positive detections.

SAR is great when it uses ground penetrating capabilities and the ability to detect metal wires and pressure plates, then have another sensor zoom in for confirmation, however cost means that SAR is still usually limited to very main roads and expensive tier 3 platforms. It appears that there is still a niche for an effective EO IED detection solution on tier 1 and 2 platforms, even it it is as backup or confirmation as they can fly lower and more regular passes.

Any views?

I linked darpa so you could search the website and the link was meant as an example of how solicitations are drafted.

With respect to imagery, I would think the fields that are currently being exploited for change detection esp automated change detection include panchromatic, color, ir, multi/hyper-spectral, polarimetric, lidar/laser besides SAR in multiple environs. There is also a lot of focus on geospatial fusion between the different image sources/types.

As this involves cutting edge tech, a lot of the performance matrix and results are IP-ed but the detection requirements are pretty standard.

Even in the field of IED detection, there are already multiple parameters in terms of requirements eg differing types of IEDs such as mines, traps, explosives, bombs, suicide bomber-type, vehicular, roadside etc.

For IED detection examples, you might want to google astamids which uses IR EO (fused with SAR).

 

[Aware]

Again thank you, you have been a valuable 'resource' and sounding board. You have given me much to consider.

Since the last post I have spent a lot of time on researching IEDs. I found a great doc by the New America Foundation that provides data and maps on IED since 2002 in Afghanistan and Pakistan, but not a lot on the way that the IEDs were detected. That is, what roadside buried IEDs look like from the surface.

I have found papers that go through all of the technologies employed and SAR, Ladar, Lidar etc all seem to work well in theory (either ground penetration or hi res 3D mapping) but several articles suggests that the high water table, salts and soil composition may hinder any radar over 1 GHz penetrating the ground which clearly all latest ones seem to to be (ie 15 GHz + range).

I also see that persistent surveillance from balloons or hovering UAVs and ground vehicle ground penetrating radars are in vogue, but like SAR, Lidar etc these are large investments, don't necessary catch perpetrators in action and are unlikely to be fitted to all arterial roads/. tracks. Yes fusion of data is going to bring all of these things together.

So I think that for any EO/IR system there are probably two main uses. Wide area surveillance (assuming the IP allows for high resolution EO detection of changes in dismount and asset location between successive fly overs. AND/or close in UAV fly-overs main roads looking specifically for changes to road surface for possible IED concealment.

The wide area option only works if the analysis can truly be a wide area swathe with low false alerts, it would be for areas that the tier 3 are not flying or at different segments of the road if say a 300 Km stretch is being survielled. The UAVs (tier 2 and tier 3) could fly opposite directions, or several tier 2 could be spaced out flying over a road, valley of interest. This is something that my clients product already does (good wide area high res analysis), I am just still researching to see that this potential ConOps is something that DoD would consider doing.

The roadside IED EO detection is another matter. All analysis programs require stable pics regardless of the sensor source. Smaller the plane or more zoomed in, the less stable. If ground penetration proves to be a problem for tier 3 UAVs SAR sensors then they are more likely to be open to a complimentary solution.

Can I get your view on this:

Articles suggest that many IEDs are non metalic/ non magnetic using home ingredients. Some are going back to trip wires because DoD is using good rf jamming equipment. If ground penetration isn't as successful as desired, then some form of surface detection (at least from the air) might still be required. Articles suggest that Lidar can detect down to 4 inches res and next generation units will be lower. They also say that the technology can detect different soil type composition - so analysis does not solely rely on a strong contrasting colour change of soil or soil being any higher than surrounding soil. This is very good, but wont be on every UAV.

If an EO/IR solution is to be used. I imagine that it will require HD sensors and/or UAVs flying low. This will not be covert, but maybe DoD would be happy for this ConOps from a deterrence factor? If the majority of the roads surveilled are dirt (no obvious pot holes dug) and soil is packed flat with little colour difference, then EO/IR detection would have to rely on heat difference signatures? This would mean still high res imagery as well as using IR or thermal sensor imagery. Since IR sensors are more often fitted as standard to tier 2 UAVs and the imagery is much clearer and high res than thermal, then IED detection using EO/IR would most likely have to rely on IR detection of different soil temp gradients, and so would mean that the image would have to be taken soon after IED deployment so that its temp does not equalise with the surrounding soil temp too quickly? During the day the subsurface soil is likely to be cooler, and at night subsurface would be warmer. Guess it depends on when the IEDs are typically placed.

OK, I get that this might be more detail than expected in a forum post, and its a specific application, but do you have any input into whether you can see either the wide area or roadside flights 'solutions' being acceptable as a supplementary IED detection method? Of course it needs to have proof of concept trials, just looking for feedback ... or perhaps its just a matter of 'build it and they will come"?

 

[weasel1962]

Re:

Just to clarify, I'm not an industry expert.

However, if you're asking my perspective, whilst IED detection is a relatively new requirement and borne out of the afghanistan experience where the conditions include air superiority (which allows unfettered use of uavs) and high levels of guerilla tactics limited by low-tech, there are already massive amount of tech and research going into this.

The challenge as you are trying to grasp is to identify existing technologies so as not to duplicate effort and focus on developing new tech. Or focus effort in enhancing existing tech. That is not something that can be covered in just a few posts.

As to IR detection of surface temp differences, that is already something that I understand is being performed.

IED detection tech is available on tier 1/2 uavs, presumably the info is channeled into a central party (possibly MI) to review and then channelled to engineer depts to clear. This includes existing wide area scan analysis, computer programs to identify possible ied locations etc. The challenge is to convince that your or your client's tech is either lower cost or more capable than existing (which DoD will compare to existing SAR and other tech capabilities).

Lower-cost identification techniques are also brought to unit level eg tier 3 uavs which have EO/IR capabilities also or ground level detection eg husky with visor 2500 ground penetrating radar.

Niitek

I think the discussion is barely scratching the surface in what tech is already out there. I think there is scope for even lower cost techniques (eg the husky costs US$m) but I suppose this is outside your scope which focusses on UAVs.

 

[gf0012-aust]

I think the discussion is barely scratching the surface in what tech is already out there. I think there is scope for even lower cost techniques (eg the husky costs US$m) but I suppose this is outside your scope which focusses on UAVs.

agree. the tech that is in the public domain barely covers what's in actual use.

as it's a highly sensitive subject I seriously doubt that anyone who is aware will be talking about capability in an open environment.

unfort/fort this means the quality of material able to be discussed is severely challenged.

 

[Aware]

agree. the tech that is in the public domain barely covers what's in actual use. As it's a highly sensitive subject I seriously doubt that anyone who is aware will be talking about capability in an open environment.

unfort/fort this means the quality of material able to be discussed is severely challenged.

I agree to an extent too. However if everything of value is so secret that no one in this forum can discuss it with any useful input, then it kind of makes this whole forum somewhat irrelevant. I was after informed opinions, not state secrets.

Anyway ... in my latest travels I have come across the 'Joint Improvised Explosives Device Defeat Organization (JIEDDO) Broad Agency Announcement (BAA)'. While some sites suggest that JIEDDO might be becoming irrelevant, this BAA pretty much describes what DoD would like to see developments accomplish in finding various forms of IED. Its closing date is Nov this year, so apart from the low and hi tech IED products/ services I have already read about, this doc probably seals my research needs for the moment.

Again thank you for everyone who has contributed.

 

[gf0012-aust]

I agree to an extent too. However if everything of value is so secret that no one in this forum can discuss it with any useful input, then it kind of makes this whole forum somewhat irrelevant. I was after informed opinions, not state secrets.

The problem that you face though is that the 3 most highly sensitive subjects in modern militaries are UDT (incl subs), Specforces and counter IED tech.

its difficult to talk about any of these matters on these partic subjects without putting some of our own careers at risk. Unfort general discussion on a lot of these areas can be meaningless.

I'm not trying to be precious or difficult, but unless you actually work in these areas its hard to appreciate the level of material that is withheld and what won't be discussed for obvious reasons . the sheer volume is significant