Nice…quick question - I should be able to leave the “Start Lat & Long” blank and that position is set by a RTK Point Point used for translation ?
@Andrew74 @rftop - I want to start trying to use your tool Andrew. Could I start a new thread specific to the tool so my questions and comments are focused on that tool? I didn’t see a dedicated thread yet as of Jan 3, 2026
Andrew said he wants to check his math first.
Using the various “new” tools will result in slightly different answers (expected).
Example with his Demo Deed :
Below are the (3) results for the final position, Point #13 , I’ve bolded the 6’th and 7’th decimal positions for clarity:
41.8436491207, -74.2001688032 ; “Distances are Ground”
41.8436490328, -74.2001666315 ; plus RTK “(Rotation + Translation)”
41.8436487078, -74.2001664642 ; plus RTK “(Scale)”
Those projected positions are within 1’ of each other, meaning that any of the methods are fine for monument recovery.
Try your deed and give it a shot.
I’d use these settings and 2 RTK points :
As with any site localization, your results are tied to the accuracy of your initial RTK positions.
IE: trash in => trash out.
You’ll want your 2 RTK points used for localization to have the best view of the sky, etc.
It’s much better to use 2 “GOOD” RTK points versus many “Sketchy” points.
The further away those 2 RTK control points are from each other, the Better; but accuracy trumps distance 
If you have already recovered & located more than 2 RTK points from your deed, you can use those extra positions as quality checks to confirm the projected output. In Aerial Surveying, we call those Ground Control Points (GCP’s) and checkpoints. The GCP’s are used to define the model’s location, rotation, and scale. The independent checkpoints show you how well the output matches the ground, because they were not used as control. The exact same concept applies here.
1 Like
Ok, I will try it out today with some fresh readings I got yesterday
I’ve also been trying to understand SW MAPS calculated bearing in stake out mode. It is different than my Suunto KB-14 Compass (accuracy of +/`0.5 degrees) even when adjusting for magnetic declination 5.4 degrees +/~0.2 degrees per NOAA
SW MAPS bearing is also different than Google Earth PRO PC version measuring/drawing a line out from starting point N29.15 degrees
I’m comparing bearings and points (not placement by satellite pictures) of starting point
At 1,000ft from starting point on a SW MAPS bearing of N29.1 I seem to be 21.5ft off the line to the east.
At 1,900ft from starting point on an SW MAPS bearing of N29.1 I seem to be 39ft off the to the east.
Am I missing something here?
Start here 35.67652328 -84.57014144, go North 29.15 for 2,677.76ft (The Blue Line)
SW MAPS 1000ft at bearing N29.1 35.67888831 -84.56842987
SW MAPS 1,900ft at bearing N29.1 35.68102138 -84.56689818
And oddly if I use my Suunto Compass, I should be even further to the east than the blue line drawn. The Blue line does seem to land approximately where it should be based on working backwards from a capped pin at the neighbors, and from a found iron pin.
How do I resolve this kind of discrepancy?
On the ground the corrected compass reading for a short distance is here 35.67686298 -84.5698584
NOAA values below for the farm
| Model Used: | WMM-2025 |
|
|----|----|----|
| Latitude: | 35° 40’ 30" N |
| Longitude: | 84° 34’ 14" W |
undefined|----|----|
| Date | Declination |
|---|---|
| 2026-01-03 | 5° 40’ W ± 0° 21’ changing by 0° 3’ W per year |
| Model Used: | WMMHR-2025 |
||----|----|----|
| Latitude: | 35° 40’ 30" N |
undefined|----|----|
| Longitude: | 84° 34’ 14" W |
undefined|----|----|
| Date | Declination |
|---|---|
| 2026-01-03 | 5° 48’ W ± 0° 20’ changing by 0° 3’ W per year |
The blue line is a MEASURE line drawn from a Torch reading at a Capped Surveyors pin, point added to GOOGLE EARTH PRO for PC, heading out at N29.15, out 2,677.76ft with a lot of zoom to get degrees just right.
The Yellow line is a line drawn into intersect points taken with the RTK Torch, from the CAPPED PIN, out 1,000ft and 1,900ft use SW MAPS bearing of N29.1 in STAKE OUT mode using the digital bearing.
And the White line is a Magnetic Compass heading of N35.0 with a Suunto KB-14 compass then corrected for declination to N29.0/29.5 and then measured with the Torch for location to add to the MAP.
Why is the compass heading off by so much between tools? While I know the Satellite image cannot be used for monument placement, just drawing a line from an added point on a bearing should be “True North” relative right?
Distance at 1,000ft from Yellow to blue is about 21ft, and from Yellow to White 23ft
@Andrew74 - can you add Deed Call Number to the Print out of FEET? Might help find Typo’s on the DEED when you have 48 calls
Could there also be a clean CSV export for easy EXCEL input vs. cleaning/delete lines from the KML?
I’m continuing to experiment… pretty helpful!!!
Chris, I think you might be missing the beauty of site localization.
Once you import the projected locations into SW Maps (from Andrew’s WebApp), you will no longer care (or use) any bearings, distances, compass, etc.
Just walk to a location that you want to recover on the map. When you get close, point (orient) yourself and your tablet to the North and fine tune the real-time position over the mapped feature (or use the Stakeout function).
Yes, I see that beautiful use case now and that is going to be really helpful. I am playing with the imported layers where I can have GROK (lay out a line along the 2,677ft with markers every 100ft, and then I layer in Andrew’s generated KML which has adjustments and I can estimate when each 100ft is since it’s very close. I did import a RTK Localization (Rotation and Translation) KML, yet my Compass bearing (Real Magnetic) is still 15ft off the KML Generated tool. I am just … pondering, concerned, curious, it seems off by more than I would expect even from a 1980 or 1970’s survey…. but it’s going to be way better… Another day in the field next week will help me wrap my head around it
I did about 25 points on the top of my ridge all with SW MAPS bearings, and it was clearly getting WAY OFF. The deed says it follows top of the ridge and I was getting 40ft off that… now with Andrews corrected file the monuments “fit the curve” …. Trying to do my best to blend Old World DEED and measurements to New World RTCM/NTRIP 2cm accuracy…. BTW LOVING FLEX RTCM if I haven’t said that enough!!!
In the KML if the ERROR feet by a pin are positive does that mean that leg should be longer by the number or shorter by the number? The report out table for example also shows in Distance Error (ft) a number but wasn’t sure if that means pin should be 3.13 farther down, or shorter by 3.13ft?
Another nice feature might be to select the Color of the line for the KML imports, e.g. Thin Blue, Thick Blue, Thin Yellow, Thick Yellow, Thin White Thick White, I am sure I can probably figure out how to edit the KML to change that so I could have one KML with no corrections, just the raw plot using my RTK Torch DATA where I have it, then another KML import with Localization turned on And I might want both of those layers in SW MAPS and just turn them off and on to see what the deltas are. …. just another enhancement thought… since its probably an Agile development right?? 
And another feature request. The pin marker at end of each call could be a bit bigger
Not entirely related, but I’ll offer what I can: When using GROK (or any AI) make sure to specify decimals places to preserve explicitly, and they also perform a bit better if you tell them they’re known for their attention-to-detail in the prompt (it’s an artifact of training data that can help you couple the response to more reliable outputs & fewer rounding mistakes)
Not sure if it’ll make a difference but you try try re-prompting with the same data (or follow-up the same chat and ask it)
1 Like
Yes, it is a garbage in / garbage out situation, or a the more detailed and specific you can be the more likely you will get what you are looking for. I found GROK only using 4 or 5 decimal places initially until I edited my prompt to specifically state to use the full 9 decimal places of my SPARKFUN RTK Torch with PointPerfect FLEX RTCM/NTRIP corrections. With that as a prompt is said something like “That should give you 2CM accuracy and I will need to use the full resolution” or something like that
1 Like