Digital Orthophotography in New York City

Orthophoto imagery is a very beneficial layer in a GIS as it provides a generic, inexpensive, and accurate base map. However, in cities with tall buildings, the creation of a seamless database of orthophoto imagery is difficult. Traditional methods for orthophoto rectification do not consider buildings. Therefore, features at street level are hidden due to building lean. Using a GIS to overlay accurate vector data on this imagery produces a confusing picture because the two information layers fail to match. In this paper, we describe a fully automated process, developed by Sanborn, that corrects these problems. The result is “true orthophotography”: seamless imagery without building lean. In particular, we present image samples from lower Manhattan that demonstrate both traditional and “true orthophotography”. TRADITIONAL DIGITAL ORTHOPHOTOGRAPHY IN BUILT UP AREAS Any aerial photograph will exhibit a characteristic known as relief displacement. Relief displacement is the geometric distortion that occurs due to elevation differences in the terrain being photographed. Objects of higher elevation, like buildings, hills, and trees, will be displaced radially outward from the center of the photograph. The greater the elevation of the object, and the further it is from the center-of-view, the greater the radial distortion. The process of ortho-rectification corrects these distortions by performing, on the image, a mathematical transformation that takes into account the shape of the terrain in the form of a digital terrain model (DTM). In traditional digital orthophotography, objects like buildings and bridges are not modeled in the DTM. Therefore, these features are distorted from their true location in the final orthophotography. This distortion shows up in the form of leaning buildings and warped bridges. In severe cases, this distortion can be aesthetically displeasing and may impact the usefulness of the orthophotography. For example, a tall building may “lean” over a street, hiding information like manholes, fire hydrants, and utility poles. Complex, multi-level freeway interchanges appear badly deformed. When a vector GIS layer is overlaid with the imagery, building outlines do not match up with the imagery representing the tops of the buildings, and vector road edges will appear as passing through buildings. In 1997, the New York Department of Environmental Protection contracted with the author’s company to create digital orthophotography for New York City. Moreover, because of the extremely tall and closely spaced buildings in Manhattan, a new approach was required for increasing the usability of the orthophotography in these areas. Figure 1 shows a traditional digital orthophoto in the lower Manhattan area. Note that the sides of the buildings are visible, and the buildings lean over and obscure streets and other features. ALTERNATIVE SOLUTIONS Described below are three possible approaches. We used a combination of all three approaches on the lower Manhattan area with excellent results. Dense aerial photography The simplest approach involves the acquisition of additional photography. More photographs of the same area means that, with this approach, we only use the central portion of each image for the digital orthophoto imagery (DOI). The central portion of the image is termed a “sweet spot”, as objects closer to the center of the photograph distort less than those at the edge. In addition, special “spot shot” aerial photography can be taken directly over complex bridges and freeway interchanges. Again, freeway interchanges falling in the center of the image show minimal distortion. The central portion of each image is manually clipped and mosaicked into the final orthophotography. In this particular project, we used three sets of photography: two sets at a scale of 1”=800’ and one at a scale of 1”=1000’.