It's over. Wreck hunters need to find a new career

[MiddenMonster]

May as well file for unemployment. This chick, Leila Character taught a computer how to recognize shipwrecks on the ocean floor from ships on the surface, and even airplanes. So join the wagonwrights and glassblowers of yesteryear, because them computers done got 'em another industry:

AI spots shipwrecks from the ocean surface ? and even from the air

 

Amazon Forum Fav 👍

Cry Of The Eagle: History And Legends Of The Cherokee Indians And Their Buried Treasure - Grab it through Amazon!

[Salvor6]

You can always turn Pirate. Keep all the profits, no taxes, no percentage to the state, much more fun.

 

[Honest Samuel]

It is no fun being in jail for not paying federal and state taxes. You should knew that I was coming.

 

[MiddenMonster]

You can always turn Pirate. Keep all the profits, no taxes, no percentage to the state, much more fun.

But you would still be competing with the AI Industrial Complex to find the wrecks in the first place. If this technology works, the droneship SS Ballard is going to be chugging around the ocean, finding all the wrecks. Then the 8-armed Robert Ballard android will be dropped into the ocean to exploit them--and every time it submerges, it gets a government grant. I could also see Elon Musk getting involved with this, which will prompt Jeff Bezos to get involved and then that's it for the little guy.

It is no fun being in jail for not paying federal and state taxes. You should knew that I was coming.

Only in the United States. In other countries you might be begging for jail time.

 

[Black Duck]

There is one place in the world you can work legally w/o any government interference, with out a contract, And GME owns the area. Under federal law.

 

[releventchair]

Fishfinder marks fish. Does not boat /land them.

And on mine , the fish all appear to be facing the same direction.....Not that it matters.

 

[crashbandicoot]

Damn computers!

 

[TRG]

Here is a link to the article - open access:

https://www.mdpi.com/2072-4292/13/9/1759

Underwater LIDAR:

https://leica-geosystems.com/en-us/...ing-underwater-terrain-with-bathymetric-lidar

The shipwreck software is open-source. Tool up folks.

 

[MiddenMonster]

The shipwreck software is open-source. Tool up folks.

Might want to invest in a good pilot, as well to cover larger areas in less time. The images from the model output are nothing short of amazing, and this could easily become the primary method of finding wrecks in the near future. But I have to wonder: How are the archaeologists going to take it when the software is used for treasure hunting instead of archaeological work? If the results hold up I could see upwards of 90% of the near-coastal wrecks being mapped in the next 15 years. There might even be a bum's rush of people and countries claiming wrecks just around the corner.

 

[ARC]

Many of you may recall my many comments on Lidar over the years... i have been watching this tech intently as it has grown for eons now.

I have always known that it was only a matter of time until its power would be harnessed via I.D. programs.

BUT... do not fret... for all is not lost...

Only what can be seen... can be found.

Anything older than 200 years is gonna be long gone and what is left is long buried in most cases... no traces visible... nothing to follow... nothing to see.

Be it computer or man.

 

[Boatlode]

Old wrecks are nothing but a pile of rocks. And a lot of those are buried under sand. I don't think finding them will be any easier.

 

[MiddenMonster]

Anything older than 200 years is gonna be long gone and what is left is long buried in most cases... no traces visible... nothing to follow... nothing to see.

Be it computer or man.

That is true--for now. But technology never stands still. In the last 10 years satellite imagery has allowed armchair archeologists to find ancient settlements, many thousands of years old while sitting in front of a computer in their parents' basement. There is no reason to believe that this shipwreck technology won't be refined in similar ways. Now it finds shipwrecks under 200 years old based on the outline of the ship. In the future it will find older shipwrecks by identifying vague shipwreck shapes based on the way the sediments or microbial growths have accumulated. It will also be combined with things like spectroscopy, which will identify deposits of metals, minerals or the remnants of decayed wood. AI will one day be able to find shipwrecks by identifying debris field patterns. I gave the title of this thread a name that was tongue in cheek, but one day, maybe a century from now it probably will be anachronistic to imagine locating shipwrecks by doing research, identifying a probable area and scanning a grid with a ship to find it.

 

[xaos]

The first step in creating the shipwreck model was to teach the computer what a shipwreck looks like. It was also important to teach the computer how to tell the difference between wrecks and the topography of the seafloor. To do this, I needed lots of examples of shipwrecks. I also needed to teach the model what the natural ocean floor looks like.

Not awe inspiring....shipwrecks come in many, many forms and sizes, and most of them are not recognizable, even up close...

Underwater LIDAR:

https://leica-geosystems.com/en-us/c...hymetric-lidar

This is not underwater LIDAR...it is the ability for a LIDAR system, with a large amount of post processing, to see through the surface of the water to where its Laser Light can penetrate.. which in reality, based on conditions, backscatter, and surface wave conditions that defract...can be very, very shallow, or a little deeper, but we all know how deep light can penetrate water...


LIDAR: Light Detection and Ranging.

LIDAR is based on the reflection from the first hard surface it "sees"...This is why these LIDAR systems for self driving cars are complete bullsheisa...

We build algorithms to post process the data to map a surface...unfortunately, the only way to test the algorithm is with the system itself. So, if it gives you the same result that you programmed it, then it is a success

This is where I have problems with this system...it is basically looking for a pre-determined result. If you find this it is this...of course it will find that, because you used that data to program the algorithm...but, if your assumption of this was not correct...then well...

Sorry, trying to put this together...you have a result from the system, you program the system to report that this particular shape or signature is this...then you test the program with that shape, and viola, it identifies it.. success?

This is the problem with building algorithms based off data...you should always get the answer that you told it to provide...

When it doesnt find what your pre-determined algorithm told it to find, you are really off...

 

[MiddenMonster]

We build algorithms to post process the data to map a surface...unfortunately, the only way to test the algorithm is with the system itself. So, if it gives you the same result that you programmed it, then it is a success

This is where I have problems with this system...it is basically looking for a pre-determined result. If you find this it is this...of course it will find that, because you used that data to program the algorithm...but, if your assumption of this was not correct...then well...

I'm not disagreeing with you, but your comments are only true at this moment in spacetime. Everything you said of this system could have been said of RADAR in 1940. But then came 1950, and 1960, etc. Now RADAR can identify specific types of aircraft and provides very good imaging. The same is going to be true for this system. The two basic components--sensors and processing (hardware and software) are going to advance at the rate all electronic technology is advancing. Machine learning has advanced light years for facial recognition and image recognition over the last decade. Now they are starting to combine that with IR and other technologies. Can we identify a specific person through a concrete wall now? No, but we're not far off. Now we could identify Laurel and Hardy through a concrete wall. We can identify a cat, or a dog now. 20 years from now there is a good chance we could identify anyone who has an image record on file. At some point, anything that leaves a baseline trace will be able to be scanned with hardware and analyzed with software millions or billions of times faster than the human eyes and brain. All that is needed is a reference library of images, spectrographs, silt piles, etc with which to compare the collected data. After that it's just a process of refining the analysis.

 

[Bad-Company]

Too much NCIS TV!

 

[SADS 669]

Whatever they find, they’ll still have to send me down to move ballast stones……

 

[ARC]

Whatever they find, they’ll still have to send me down to move ballast stones……

Which is where the REAL work would begin. heh

 

[pepperj]

Sitting across the dinner table QM ll in 2006/7 with a fellow from Lockheed Martin, the conversation turned imagery and technology.

"All I can tell you Jim is this-what ever you can see today, is what we saw 20 yrs ago"

Now fast forward to 2021-what do they see-to compared to what is available for the general public.

I would love it if the LIDAR system of mapping coverage was available for my hunting area.......

 

[Blak bart]

Whatever they find, they?ll still have to send me down to move ballast stones??

Im right there with ya SADS.....finding and recovering are 2 very different things....both of which can be very difficult !! Now we have fancy technology but you still have to pick which wreck to salvage. I have many wrecks that I have found over the years.....only a few would be worthy of salvage and warrant a full fledged salvage op. Unless the wreck has some exceptional history or a major treasure on it no one wants to bother.....including our archeologist friends !! You have to find the one worthy of a funded project. There are several wrecks here in the keys that still have treasure on or scattered around them. Salvors moved on because the treasure was scattered over a wide area and difficult to recover.....it was easier to go to wreck sites that were more lucrative. Technology helps no doubt about it.....there is no guarantee of treasure recovery !! I have seen very advanced technology, and huge operations with many backers and lots of funding sail away empty handed. On the other hand....skin divers such as myself have gone out with little more than a mask snorkel and fins, and come back with silver and gold coins !! Like captain Ron says...if its gonna happen it'll happen out there !! Im not giving up yet !! It'll be years before technology like this is applied and used effectively on a big enough scale and all around the world to really be useful !!

 

[Blak bart]

I need a magnet that only sticks to gold.....I could use it for a few years and then sell the technology for big bucks....after I have a vault full of gold !!