About the Project
How it started We started on June 3rd 2004, with little knowledge about the internship, or our co-workers. Software We are using PCI Geomatica V9.1 which contains several applications. We used Focus and OrthoEngine. OrthoEngine helped us with tools for extracting DEMs from aerial photos and satellite images. The Process We scanned the aerial photographs and organized them by their respective years. We had meetings with Dr Jack Cothren, to show us how to use ortho engine and explain the theory behind the process. Our first step was geo-referencing, which gives the photographs real-world coordinates, using ground control points (GCPs) and positions the image (camera) in the world. Our second step was Orthorectification, which establishes a constant scale photo and negates the effects of the terrain.
The Process with Ortho Engine
When setting up the OrthoEngine project for georeferencing and orthorectifying, we had to choose the best math modeling method, depending on how much information was available about the photograph. We used Aerial Photography, Rational Functions, and Polynomial. When using Aerial Photography we had to set projections, the Output Projection and the GCP Projection. Then we had to set Standard Aerial Camera Calibration Information. This included General Camera Parameters, Radial Lens Distortion, Decentering Distortion and the Fiducial Marks. In Polynomial and the Rational Functions this was not required. Now we get to are processing steps: Project, Data Input only in (Aerial Photography), GCP/TP Collection, the Ortho Generation, and the Reports. GCP/PT Collection is where we collected the GCPs. In this we had to name point id and load a Geocoded Image and select a DEM and extract elevation. The GCPs were double checked for accuracy. Once we were satisfied with the accuracy of the model, we generated the ortho-rectified image, using the “nearest neighbor” resampling method.
Once the ortho-generation was complete, we opened the image in Focus to check the accuracy of the ortho-generation, by overlaying the streets and a 2003 Quickbird image of Fayetteville. The roads and buildings had to be within 10 meters of their true location before we were satisfied (rather impressive, given the lack of information available about the photographs). If the images did not meet our standards we rechecked the GCPs. The images were clipped to minimize the excess boundary, and exported as Geotiffs with world files so that they could be used in GIS software.
The Conclusion After millions of notes and loads of coffee and insane pressure we made it. We are disappointed that this took a long time to do but we are very proud of our accomplishments. The city has the ortho-rectified images available online (soon) so please visit their site www.faygis.org or contact the GIS department at the City of Fayetteville. We have created a couple of videos that will allow you to explore Fayetteville as it has changed through time, and an online mapping page is in the works.
