Adjusting a boundary shapefile based on GPS data points

765
8
09-23-2013 12:56 PM
GabeRoxby
New Contributor
I have several shapefiles representing boundaries of properties.  The length of the sides of these polygons is accurate, taken from surveys.  However, their location is not always accurate.  When visiting these properties, I took accurate GPS information at some of the boundary corners.  What I'd like to do is be able to match up these GPS points with the corners of the boundary shapefiles and have GIS adjust the shapefile. 

My first thought was to use Georeferencing, but I'm not sure that works for shapefiles which already have spatial information.
I then tried to use the Spatial Adjustment toolbar, but I'm not sure it has the settings I need. It seems like I want to spatially transform the shapefile (as opposed to edge matching or rubber sheeting), but I only want to rotate and translate the image, without doing any differential scaling or skewing.

I'm sure I'm missing something here, but can't figure it out.  Any thoughts?

Thanks in advance.
Tags (2)
0 Kudos
8 Replies
DavidParker4
New Contributor II
g

you did not mention the directions.  accuracy and precision? how accurate is the gps and what precision?  Was the survey entered on a map set to a geo coordinate system that was good & correct for the vicinity of the properties?  If you gps to within a meter is it close enough or do you need to be within a 1/4 inch?

dave
0 Kudos
GabeRoxby
New Contributor
The GPS points were taken as 2 minute averages, and the GPS indicates that most have an estimated accuracy of 5-11 feet.  I believe the survey directions are accurate, but am not sure specifically.  The bottom line is that the location of the corners (easily identifiable on the ground) is much too far from their location in GIS, sometimes up to several hundred feet.  I just want to make the boundary shape in GIS match up more closely to what's on the ground in reality.  This ground reality is marked as accurately as I can by the 2-minute-averaged GPS points.

g

you did not mention the directions.  accuracy and precision? how accurate is the gps and what precision?  Was the survey entered on a map set to a geo coordinate system that was good & correct for the vicinity of the properties?  If you gps to within a meter is it close enough or do you need to be within a 1/4 inch?

dave
0 Kudos
JoshuaCole
New Contributor
If the polygons and points disagree by hundreds of feet that might be a projection issue.  Could you share the projection information for both the polygon parcel and GPS point layers?  You might also want to double check that both layers are in either feet or meters and that the projection is the same as the data frame's projection.  -Josh
0 Kudos
GabeRoxby
New Contributor
The projection of the polygon is in Transverse Mercator, using the State Plane coordinate system.  The GPS points were taken in Transverse Mercator.  Both are in feet.  The data frame's projection is also Transverse Mercator, using State Plane coordinate system.

Most are not off by 100s of feet, and they're not uniformly off (i.e. all are 150 feet too far north, etc.)  The boundary polygon was developed years ago using COGO from survey information.  The shape itself is accurate, but its location is not.  Perhaps the best way is to manually rotate or move the shape so it seems to best fit the GPS data?  I was hoping for a GIS process that minimized the errors, but perhaps that doesn't exist?

If the polygons and points disagree by hundreds of feet that might be a projection issue.  Could you share the projection information for both the polygon parcel and GPS point layers?  You might also want to double check that both layers are in either feet or meters and that the projection is the same as the data frame's projection.  -Josh
0 Kudos
JoshuaCole
New Contributor
OK, probably not a projection issue.  Have you tried Snap (editing)?  Or maybe Integrate, though this one doesn't see quite right for you.   I found these by searching 'snap' in my geoprocessing tools.
0 Kudos
JohnSobetzer
Frequent Contributor
You have several options it seems to me, and all of them will involve some fudging.  I'd use copies.

The first problem is that surveys differ and can differ a fair amount.  So abutting surveys likely won't match.  The second problem is the GPS can vary depending on crown coverage, satellites, etc.  Third changes in one area have ripple effects through a whole collection.

One approach would be to move and rotate the individual shapes manually to best match their GPS, then assess how messy the result is overall. If it is pretty darn good you could snap points together or even integrate it.  Put it likely would be messy.

A second would be to select all the shapes, then move and rotate the collection to best match up with the GPS.  I'd start with this.

A third would be to do a spatial transformation.  They vary in how much they change the shapes.  You might try a similarity transform.  See http://help.arcgis.com/en/arcgisdesktop/10.0/help/index.html#/About_spatial_adjustment_transformatio...
0 Kudos
GabeRoxby
New Contributor
Yes, this is what I'm hoping to do.  Move and rotate shapes individually to best match GPS points.  So you're saying manually moving them is the thing to do.  There is no tool that I can set multiple control points (i.e. match up GPS points taken at corners with their corresponding corners on the boundary polygon) and allow GIS to find the most accurate orientation?  The Spatial Adjustment tool does this, but I can't find a way to tell it NOT to distort, enlarge or shrink the shapes.  I only want them to move and rotate.

Thanks for your help.



One approach would be to move and rotate the individual shapes manually to best match their GPS, then assess how messy the result is overall. If it is pretty darn good you could snap points together or even integrate it.  Put it likely would be messy.

0 Kudos
TOMKONG
Occasional Contributor II
Yes, this is what I'm hoping to do.  Move and rotate shapes individually to best match GPS points.  So you're saying manually moving them is the thing to do.  There is no tool that I can set multiple control points (i.e. match up GPS points taken at corners with their corresponding corners on the boundary polygon) and allow GIS to find the most accurate orientation?  The Spatial Adjustment tool does this, but I can't find a way to tell it NOT to distort, enlarge or shrink the shapes.  I only want them to move and rotate.

Thanks for your help.


It's a interesting issue.
Where is your source parcel data from surveys? How was it loaded (converted) to GIS data?
If the process is correct, it shouldn't have that issue you mentioned above. You should check it first.
0 Kudos