Satellite Ai For Damage Assessment ?

https://www.c4isrnet.com/intel-geoi...automate-damage-assessments-after-a-disaster/


Good morning all,

Per ...

I can accept satellite AI for forest fires. Other roles, such as hurricanes - polluted flood waters inside structures cannot be assessed from up there. I'm sure the contractors will develop all kinds of algorithms and probably charts - meanwhile, boots, rubber, sturdy, worn by the experienced and knowledgeable, will provide the assessment.

How can satellite imagry determine operational roads ? Looks can be deceiving. I'd also prefer airfields to be inspected by "boots on the ground" - - unless it's accepted by the aviators and managers that the aircraft is enroute to just a "field".

 

Regardless of the capability of the pattern recognition algorithm, any overhead platform (satellite or aerial) should be seen as a part of an integrated system rather than as stand alone system.

 

Good afternoon Varuna,

That makes my point.

The satellite programs get the big public funds. The field and amphib responder teams survive on scraps and personal responder funding.

 

The imagery doesn't determine anything.
Humans look at the imagery and make determinations.

AI can be used to relay accurate assessments too though.
A washed out road is easy to see.

Even a computer can tell Hwy 12 is out of service.



That was in the link you posted also:

 

Good afternoon Snyper,

Exactly !

A human makes an interpretation of the satellite product.

Yet, ref your above last para re the link I posted, still distills down to the analysts' work as being too dangerous to rely on regardless of the speed sent to the on the scene responders.

The road might indeed be passable but those traffic barrels in the sat photo are really commingled lots of drums of oxidizers and corrosives. The drums are damaged.

The boots on the ground responders will be replaced by the high tech IT when the military goes full push-button warfare. We've got a short wait - unless there are contractor delays.

 

I can't tell for sure the US situation on the matter, however just like other country it usually due to the need of the policy / decision maker to deliver some kind of report to the public regarding the scale of the disaster.

 

Here are some example between what you get from overhead in this case aerial asset compare to ground asset

Overhead (aerial)



Ground




Overhead (aerial)



Ground

 

Good afternoon Varuna,

True, here also, the overall political situation gets the focus. The public must be keep briefed and getting responders to the scene is just too slow. My point is still that the responders get little attention for their assignments.

Those are great pictures ! Are they from the Surabaya area ?

 

It's surprised me that there is insufficient budget for US first responder compare to the supporting asset (overhead imagery)

They were from Palu Earthquake & Tsunami back in Sept 28th, 2018. The earthquake trigger soil liquefaction and undersea avalanche thus triggering tsunami.

The disaster is a great example of "slow respond" but in this case it wasn't due to some lack of political will or budget (we have sufficient budget, material, and manpower for it). But it was due to the destruction of access (port, runway, and road) to the affected region. The only way to reach the region was either by V/STOL capable plane or by amphibious operation.

These what happen to the roads ;

 

Good evening Varuna,

The satellite programs are Federal, that is the national government.

Much of the responder organizations (but not all) are state-level or city, county. The States cannot - properly - handle the costs.

Again, great pictures !

 

Why would anyone think that without more data?
Who said anything about "traffic barrels"?

That's a Terminator movie script.

 

For now (present days) that maybe the case, however after the year 2040 that so called Terminator script has a real possibility to become reality

 

I prefer to stick to "now" when it comes to most things.

I also prefer "probabilities" over "possibilities".

I'm not worried about Skynet being activated.

 

I was originally intend to use "inevitable" rather than "possibilities", but I thought it was too dramatic.

The coming of a strong AI / general purpose AI is inevitable, and I'm using the year 2040 based upon the technological progression in the related field particularly machine learning, neural network, quantum computing, etc. Pattern based learning AI are already around us and running so many aspect of our daily life (finance, commerce, political campaign, public utilities, etc), and there is no stopping of people like me who always seek ways to further improve those AI for our benefits.

 

The development and advancement of the technology is "inevitable".

The idea of it totally replacing humans for most functions is still in the realm of science fiction.

I'm not opposed to that at all.
I'm simply saying I don't foresee it totally replacing humans for most things.

Other's opinions might vary.
That's fine too.

 

Eliminating any human factor in any job is not technically viable option within any foreseeable future either. Even in highly automated environment there is always the need for human factor for supervising, managerial, and decision making role. However the type & complexity of task that could be handle by AI / robotic will only increase as time goes.

 

Even for analysing forest fire there is still significant time delay between data capture, data processing, aerial / ground verification, disposition.

Example, this particular data is about 6 hours old, and it require some guess work to determine where or how big the fire on the ground thus still needing aerial asset (drone)

http://asmc.asean.org/home/

 

That's what I said to begin with.