In early June of 2011, the Autodesk software company released “Project Photofly” in its Labs division.  Following significant feedback from a variety of fields interested in 3-D modeling, Autodesk decided to turn their lab project into a fully developed, marketable, and operational software package called “123D Catch” that is currently in Beta testing.  This software generates a point cloud based 3-D model from nothing more than a series of digital photographs taken 360° around any given object.

Around the same time that Autodesk decided to develop their software, research began into the local Pittsburgh, Pennsylvania legend of a B-25 Mitchell bomber that crashed into the Monongahela River on January 31, 1956 and was never seen again.  A recent 2008 river survey conducted by the B-25 Recovery Group and various other public and private agencies turned up some compelling evidence as to the location of the missing bomber.  This evidence comes mostly in the form of ROV video footage of what appears to be a propeller and possible remnants of aircraft landing gear. Diver testimonials also refer to propeller and wing-like objects present at the site in question.  The search was called off, however, due to dangerous diving conditions. 

The dangerous conditions and hazardous debris littering the Monongahela River make sending inspection divers into the river a last resort.  Photographic images and video footage obtained by an ROV would help to alleviate this risk to divers and may, in fact, provide enough information to identify the wreckage. 

However, the authors would like to proceed further and create a 3-D model of the wreckage.  This model could be useful for a number of applications including the development of possible excavation plans, the creation of museum displays and exhibits, and possibly to help determine what has happened to the site since the time of deposition.

The development of Autodesk’s “123D Catch” software makes it possible to gather 3-D data remotely due to the simplicity and ease of operation of the necessary hardware.  The dynamic nature of the Monongahela River and similar environments, however, create an obstacle to the proper positioning of an ROV for the “123D Catch” software to function properly using still photography.  Theoretically, it should be possible to gather the necessary images using an HD video camera fly-around.  This method provides for much simpler data collection, since the video camera eye is always open.

The authors, therefore, tested the ability of “123D Catch” to create 3-D models using frame grabs from HD video footage of objects on dry land.  The technique is then tested using footage gathered from a controlled underwater survey of a known object.

If successful, this method shows tremendous promise for the future of remote 3-D imaging. The technique could be used to gather 3-D data from a variety of settings that pose hazards to humans and certainly should not be limited exclusively to imaging archaeological material.