I have a large (equivalent to size of Ireland) line feature class (fc) depicting a shoreline, which I am editing. The shoreline is in Alaska State Plane Zone 1 (meters), NAD83. The NAD83 realization is closest to NAD83(1986) or NAD83(CORS96). The shoreline is in an ArcMap document. The dataframe is set to Alaska State Plane Zone 1 NAD83 [EPSG26931]. I am using ArcMap v. 10.1 on Windows 7.
I have two DigitalGlobe (DG) map services in the ArcMap document:
1) DigitalGlobe Web Map Service:Imagery, which is served in WGS84 [EPSG: 4326], and comes with an option to alternatively receive the image in WGS84 Web Mercator (auxiliary sphere) [EPSG: 3857].
2) DigitalGlobe:ImageryTileService, which is served in WGS84 World Mercator [EPSG: 3395]
I have no idea which realization of WGS84 is represented in these orthos.
I would like to be able to use the DG ortho-imagery for digitizing new shoreline information and to update existing lines in the NAD83 shoreline fc. The problem is getting the 3 data sources to line up. For the dataframe properties I have specified the WGS_1984_(ITRF00)_To_NAD_1983 transformation.
--I have some lines in the NAD83 shoreline digitized from in-house orthophotos that are also in State Plane AK Zone 1, NAD83. These features are about 8 meters off from the DG Imagery and 10 meters off from the DG TileService. However this could be explained by spatial inaccuracies from the orthorectification.
--The DG Imagery and DG TileService orthos are about 12 meters off from each other. It seems I should be able to get these 2 sources to align with each other.
1) Is there a transformation to align two images in WGS84 [EPSG 4326] & WGS84 World Mercator [EPSG: 3395]?
2) Can I align EPSG 4326 & 3395 with NAD83?
Many thanks to anyone who can help.
Thats a good one - down in the sub 10 meters for relative comparison, so it could be an issue of "WGS84/NAD83", or at least, that 1.3 meters or so could be playing a part. This can be distilled to a datum, not CS or PCS issue.
Your stuck without TRUTH on the ground with VIO (visible Identifiable Objects). GLBA has about 10 that I trust. Would that help? By settling on "truth, hopefully seeing the same VIO on all three datasets and altering transformations to fit, can you resolve the issue.
Call me. I'll try to help.