New Geo References Ahead

Over a portion of my career I performed a lot of aerial mapping. Processing the geolocations involved a lot of conversion between different datums and projections. No big deal if you know the mathematical equations for the conversion. I suspect that GPS receivers will need a firmware update to handle the new reference. I would speculate that most GPS receivers will probably continue to use WGS84.
Mark

That was an interesting article - thanks for posting. I've worked a lot with aerial imagery myself in recent years. While it's one of my smaller maps, this one involved stitching together 1970 imagery from 6 different sources, one of which involved manually georeferencing and cropping almost 400 raw photographic plates. That took me weeks!

https://boydsmaps.com/mapinfo/pines1970.html

But recently, I realized that I was thinking about this as a "human" while today, computers can do things so much faster - but you have to step back and consider the problem in a new light. The next time I do a project like this, I will just take the raw imagery and use software to match features (the raw plates have considerable overlap). To do that, it makes no difference what the coordinates or projection is, the software just looks for matching features and stretches/rotates things until they fit. That is very different from my usual approach of individually georeferencing each plate.

And it turns out, there's open source software designed for stitching together drone photography that looks pretty good. Once all the images are matched up into one big map, then you georeference the whole thing in one shot, and I beleive there's software for that too. AI can do all of this by itself actually, but the issue would be the cost, because of the amount of data. But the alternative is to use an AI-guided approach, where it recommends the software and tells you how to configure it. That should be pretty straightforward and not involve any big computation expenses (because I will do the processing on my own computer).

Well, that's the theory at least. I hope to put it to the test in the future, because I have more raw aerial imagery that I'd like to process.

. . .I worked in the geophysical industry guiding 65 to 200+ feet sized vessels using proprietary radio positioning systems and occasional satellite. These systems were based upon post-WWII systems that pre-dated Loran and the eventual transition to GPS. Like the technology upon which it was founded, the company faded away within the decade after I moved on.

One of my company's systems produced a concentric wave signal and the other one was hyperbolic as cross checks and finer accuracy where one had an advantage over the other.

What datum(s) the recorded readings were based upon is long forgotten, if I ever knew. Maps of the surveyed areas were produced afterward from coordinates recorded. The goal during the survey was typically guiding the vessels along pre-calculated coordinates forming parallel lines and a couple of crosses that intersected the parallel lines. Combined with seismic data or side scan sonar plots, the clients typically used the plots to determine lease sale bids for drilling wells or to document that nothing got left behind on the ocean floor after a rig was moved off a drill or production site. Sometimes, it was to bring divers to inspect or repair a pipeline valve or other submerged object.

I had no idea at that time how my future would continue in the various fields that followed, but the one common thread was it always involved high tech electronics. As I read the article, I found the dates of standardized references that came along after my days doing navigation particularly notable while bringing back some dusty memories and another opportunity to marvel how technology continues rapidly advancing and changing our norms.

I like to read technical stuff. That was a good article.

Granted, it's a bit of a long read, but am happy I'm not the only one who enjoys a deep dive into a subject of interest.

We used a technique called direct georegistration to determine the latitude and longitude of each image pixel. With direct georegistration, the precise position and attitude (roll, pitch, and heading) of the platform is measured for every image. Each pixel in the image is ray traced to the ground. The latitude and longitude of the pixel is determined by the point where the ray intersects a DEM (Digital Elevation Map), producing an orthorectified image.. The INS (Applaxix POS-AV, price ~$300K) we used required post processing to provide cm position accuracy (This was prior to development of RTK). We were unable to warp images to control points because the imagers we used were "push broom imagers". With push broom imagers, each image frame from the sensor is an independent cross-track line in the final image. The in-track image pixels are created by using the forward motion of the aircraft to build up a two dimensional image.

Yeah, digital changes everything. But I'm intested in legacy imagery, taken with film cameras such as the NHAP (National High Altitude Photography Program) in the 80's and NAPP (National Aerial Photography Program) in the 90's. Then there's a nice archive of tri-state 1970's imagery from the Delaware Valley Regional Planning Authority (DVRPA).... and I'm sure there are a variety of other alphabet-soup organizations out there with imagery, just waiting to be georeferenced!.

Here's some cool open source software I want to check out when I have the time (but too busy with other things now...)

https://opendronemap.org

Hmmm, I have a few intersections defined on my nuvi 2689 using latitude and longitude. Most are on unpaved undedicated roads in upper New York state. One in particular is at the driveway to our hunting lodge where it intersects an unpaved road. The 2689 gets me there flawlessly, even in an occasional blizzard when using a snowmobile.

I guess I’ll have to be careful in the future or I’ll end up missing the drive and end up in barb wire!

Given the last time I rode was back in the mid-70s in the U.P. of MI, my knees would probably get their revenge shortly afterward, but sure would enjoy getting to race a Yamaha 433 over some freshly fallen powder once again. Even at a maximum revision around 4 meters, I expect you'll spot that entrance readily, although it'll be interesting to discover how this new datum affects real world users.

The rollout of the new georeferencing system will not change how consumer-grade GPS receivers work. The updates being introduced by the National Geodetic Survey (NGS) to the Modernized National Spatial Reference System (NSRS)—replacing legacy datums like NAD 83 and NAVD 88 with modern reference frames like NATRF2022 and NAPGD2022—primarily impact backend data processing, mapping, and high-precision survey equipment. The transition leaves consumer devices unaffected due to a few key technical realities:

1. Different Native Reference Frames

Consumer-grade GPS receivers (found in smartphones, fitness trackers, and standard car navigation systems) determine location using global reference frames like WGS 84 or ITRF. The upcoming NSRS modernization updates regional North American datums. Because consumer devices do not natively calculate positions based on these regional datums, their basic positioning logic remains unchanged.

2. The Built-In Error Margin of Consumer Hardware

Standard consumer GPS chips typically feature a standalone accuracy range of 3 to 5 meters (10 to 15 feet) under normal conditions. The shift in the new georeferencing system corrects regional tectonic plate warping and elevation tilts by roughly 1 to 4 meters. Because this shift falls completely within the existing noise and margin of error for consumer hardware, everyday users will not notice a difference.

3. Background Software Handling

For consumer apps that rely on precise alignment (such as Google Maps or Apple Maps matching your phone's GPS pin to a road layout), the heavy lifting is handled entirely on the cloud or app software level. Map service providers update their baseline imagery and GIS layers to match the new coordinates automatically, ensuring the user experience stays seamless.

I've never known how to read that. Does it mean e.g. ±5 meters with a 90% confidence interval?

Some interesting data here. I saved a link to this site long ago, and most of my old links from back then no longer work. So, I was surprised to find that this one still did work, and even includes some recent devices...

https://www.fs.usda.gov/database/gps/mtdcrept/accuracy/index...

Yes. The confidence level can approach 95% with newer GPSr's under ideal conditions. Some devices display it as EPE (estimated position error), and can often be much less than 5 meters. It depends largely on the satellite horizon at any particular location.

. . .got a page saying the "Request was blocked." Guessing if it worked before, possibly due to a browser incompatibility, but seems out of the ordinary given the government site.

Weird. Just tried again and it works fine for me on the Mac in both Safari and Chrome. Also works on Edge in Windows for me. Maybe it was down when you tried? Or maybe you have some kind of anti-virus that blocks it?

Thank you. That was of genuine interest.

I looked at the data for GPSMAP 64 since I still occasionally use one. Right away I got confused. With WAAS turned off the device got more accurate when the number of samples increased. That part I understand. With WAAS turned on the reverse happened! I don't grasp that at all. Maybe that should tell me that I don't really grasp WAAS. I thought I did.

Then, then(!) the table of accuracies didn't have units! Meters, feet?? ± or not??

There is a note saying " For more information on NSSDA testing check out the following link: NSSDA Standards ". Following my nose it says: " Positional accuracy values shall be reported in ground distances. Metric units shall be used when
the dataset coordinates are in meters. Feet shall be used when the dataset coordinates are in feet. "

I hope that I don't appear to be blaming the messenger. I appreciate what you've done. Again,,, thanks.

I also wondered about what units they were using. This is a topic we discussed at GPSReview "back in the day", you might be able to still find it at archive.org?

I did a review of the Garmin GLO at GPSTrackLog way back in 2013. That was when Rich was still running the site and it was great. He sold it to some travel agency (?) long ago that basically just turned it into an ad platform. It has gotten very spammy, so use this link at your own risk! But I did some tests of the GLO vs my Montana 600 and 60csx. Some of the side-by-side comparisions are interesting. Looks like the whole article is still there, but just beware the spam minefield.

Scroll way down to get to the track and waypoint tests:

https://gpstracklog.com/2013/04/garmin-glo-review.html