Corner points when creating boundary

To have a button to create a corner point in stead of recording a few meters of line.
Don’t know if it work because you’re position should be steady when standing still.

Is it possible to get a button for such cornerpoints and to connect the points to a boundary.

Would you be walking and asking AOG to just record the corners, or would you be typing them in? We can already pick four corners in google earth and import to AOG.

What I think you are asking is instead of moving, just pick and record a point here. Is that correct?

Yes, it is correct. But it is only for the case you use dualheading, when you go with the Tractor.

With google earth the deviation is about 30 cm.

My idea is to go with one antenna on a stick.
It would be the accurate way.

Richard and Franz, this is indeed a thing. If you go surveying in the field, you want to do it only once. With the risk of fields suddenly not being compatible with newer versions of Agopengps, you might consider creating a survey outside of Agopengps for once and for all. The missing link so far is how to go from a survey with given coordinates (CSV, shapefile, …) to a KML file that can be imported into Agopengps, hopefully we can find a way. See also: KML creation outside Google Earth

Ok, I understand.

  1. How often will be changed the file of the boundary in future?
  2. When we talk about converting boundary from anywhere, the accurate is about 30 cm+ because that’s normal one pixel on the screen.
    Now, when we talk about Uturn, section control and … with mm accurate, how is this possible?

The only really accurate way antenna on a stick and to write the coordinates on a paper as backup. It should be found a way to
Create an accurate boundary, which is reproducible.

The problem with AOG and this, is that you always have to “take the tractor with you”. To keep the tractor from spinning it will not move until you have moved far enough to give an accurate direction.
If you can get the cords from an outside source, you can edit the kml file easy enough.
You could also build a boundary from a txt file easy enough.
But, I don’t know how to suspend the tractor to walk to a corner.
When importing the kml out of Google earth and then overlaying the built kml file into google earth, there was an offset of 1 meter or so. Where’s the error? I don’t know. I assumed in rounding.
Let’s try this. Export a kml file from Google earth. Open file and edit your cords. Then create in AOG. Then check it for accuracy.
If we can verify that this works as precise as you like, then we could build further.
Number of decimal places is important. More is better. But can’t be more than AOG can handle.

Now I put a antenna in the corner and get the position from Agopengps. Then I wrote it into the kml.
When we take the position of then antenna the heading is
Not important.
I thought there is a easier way.

Does that give you an accurate point to go back to? I mean, can you walk back to that corner with AOG and it be right?

I find that no matter what I put in the kml file. The output kml file is off by half a meter or so. I can put the exact field lat and long into the Google earth kml, rebuild the boundaries and I’m still off the same amount in the same direction. Just now my field has shifted away again.
Once the boundary is created, all the points are spot on and do not change. Purple is kml out of AOG, red is kml into AOG.

IMG_20191218_114415

That is exactly what I mean with accurately of a kml.

I take the coordinates from the main screen.

If you just could have a button in the record boundary form that does create a corner point which is including a + or - offset that you enter, let’s say an offset from middle of your tractor (antenna) to the outside of your tires. That could be left side (-x.xx ) or right side (+x.xx) of your tractor.
Hit the button and you have your corner point.

There may be something going on with the import kml.
I built a routine that drops a flag at the four corners using the ge kml. Then changed it to the first point on the boundary, definitely off.

IMG_20191218_133745 IMG_20191218_134100
IMG_20191218_134115

To Richardklasens
That’s only possible if you work with dualheading. When you stop the tractor with one antenna, you have no heading - so where is left or right. With one antenna you can only use the position of this antenna.

To kentstuff
You have also the divergence of your base fix to google earth.
So it isn’t possible to get a really accurate boundary.

You are going to need to build the boundary from the easting and northing. Not lat and long. There could be a button for that I think.

Ok, thank’s.

When i drive to the corner of my field and stop, i could make a corner point if there was a button.
It’s really stable at my PC and setup with just one antenna.

I use a Tinkerforge Brick 2.0 as Heading device

Yes, then it’s ok.
For me it is no problem, I go with dualheading.

You actually don’t need a heading when making the boundary points, just the x,y position. Not sure how many digits resolution are stored in the kml, maybe they are chopped off too much.

I’ve never relied on google earth for boundaries, it can be anywhere from the right place (often just airports or very densely populated cities) to 60m off where it should be. Not a misprint, that’s 60m.

@Franz - Convergence angle (CA) - azimuth error is complicated to spin everything (unwind) all the points before getting back to the original lat lon. I do know the lat lon to UTM conversion functions work well, the conversion ends up exactly in the right place again. However adding in convergence rotation back in and out may affect position i’m not aware of.

As usual, all this stuff is grossly complicated. I asked on a few GPS/GIS forums about how we do the CA rotation matrix on the points and most replies have been blank stares, or the usual response “Well if it works then you are doing it right”. Not exactly helpful. But i do know that going straight north via utm guidance, it’s 0 degrees exactly whereas previously without CA and being on the edge of a zone, it’s way out.

Not sure any of what i wrote here is helpful, but if there is a problem, we need to figure out the why.

1 Like

It would seem something is correct at least. Set the sim to 53, -111 and set a pin, set GE to read in utm and the values are within a cm. 53, -11 is zone 12 500,000 northing and 5872270.527 which is exactly what both google earth and aog and the exported pin to GE all say. Set units to utm in GE to see them instead of lat lon.

In the image you can see it all lines up

2019-12-18 (5)