There are some different two approaches that one can take to provide the desired surface from bridges for use when rectifying orthophotos. You may even wish to use a combination of the them.
- One method would be to comb through your dataset using the QAQC Toolbar to methodically look for bridges and then use the manual classification tools to classify those bridge decks into class 17. Once you have your bridges in class 17 then you can use a combination of the ground and bridge deck classes to generate model key points which could then be used to support the ortho.
- A second method would be to locate the bridges, again using the same technique, but instead of classifying points, simply use the breakline tools to draw breaklines to represent the bridge deck. For most bridges a simple two line model of the width of the bridge and conflated to the elevation of the bridge LIDAR returns should suffice. The model key points then used in conjunction with the breaklines would formulate the DEM for use in generating the orthos. The advantage of the breaklines would be that you would likely end up with smoother edges to your bridges with less effort than the manual classification method.
- Perhaps a third method, which would only apply if you have planimetric polygons for the bridge structures would be to conflate these to the LIDAR data in order to avoid needing to classify or digitize breaklines. You would just need to be careful of the conflation method that you elect to use in order to ensure the polygons end up at the elevation that best represents the bridge deck.
In either scenario you may be able to speed up the process if you have existing indicators of the locations of the bridges in which case you could use the coordinates from these in the control point toolbar to drive you to the bridge locations to then manually classify or digitize the breaklines to add the bridges back into the DEM for use in the orthophoto generation process.
