Image rectification for recompiling and digitizing soil survey maps

ABSTRACT:

The U.S. Department of Agriculture Natural Resources Conservation Service (USDA NRCS) is placing soil survey data in digital format. To accomplish this mission, existing soil survey information must be digitized. Map scale differences and publication of older surveys on non-ortho aerial photography complicates this task. For effective use, soil survey maps need to be recompiled and digitized to an ortho-photo base at 1:24,000. This paper describes an image matching technique to facilitate the recompilation and digitization of published soil survey reports. In addition, the accuracy of using a polynomial equation versus a triangle-based finite element approach for rectifying non-ortho aerial photographs to digital ortho-photo quadrangles (DOQs) was evaluated. It was hypothesized that the triangle-based method would be more accurate if a large number of ground control points (GCPs) were used. However, results showed that accuracy differences were minimal and that the critical factor was the quality, not quantity, of GCPs. Our image matching technique is a cost effective digitizing approach for regions of low to moderate relief.