Collecting & combining GCP during creation of Mosaic Datasets

Discussion created by hlzhang525 on Apr 10, 2011
Latest reply on Jun 14, 2013 by hlzhang1022

In ArcGIS 10, we can do rectifcation with RPC or ortho-rectifcation with RPC & DEM from Raw Bundle of imagery (IKONOS, GeoEye, QuickBird, OrbView, etc). But, we couldn't find any efficent ways to collect and combine GCP from references or surveying data to improve accuaracy, during creation of Mosaic Datasets.

Does anyone know if we have any way to do so in ArcGIS 10? or

Does anyone know if ESRI has any plan to provide such kinds of capability in ArcGIS 10.1 or later?


For engineering-grade accuracy of processed images, 2-3 pixels are allowable in shift (horizontal), corresponding to historical GIS & imagery references or surveying data [those references can be used to collect ground control points (GCP)].

As we know, for example, with RPC and very-high-resolution DEM, GeoEye-1 can offer up to five-meter (10-pixel) geo-location accuracy, which means that customers can map natural and man-made features to within five meters (about sixteen feet) of their actual location on the surface of the Earth. So far, no other commercial imaging system can do this. [Of course, only using RPC without DEM, GeoEye-1 geolocation accuracy becomes poorer, even up to 10-meter (20-pixel) in shift...]

Link to the following research for more :
1. http://www.asprs.org/publications/pers/2009journal/june/feature.pdf
2. http://mcgsc.usgs.gov/publications/Kaintuck_vaughn_.pdf

So, without GCP's integration during rectification or ortho-rectification, the spatial accuracies of finally-processed Mosaic Datasets are usually not enough to meet the basic requirements for majority corporate applications.