AUV Camera Capabilities for Deep-Water Archaeology

Autonomous Underwater Vehicles (AUVs) are built for a variety of purposes and come in many shapes and sizes with near limitless combinations of sensors and payloads.  Some are built solely for oceanographic uses, collecting water column data salinity, dissolved oxygen content, etc., while commercial survey AUVs are designed to collect geophysical (e.g. side scan sonar or seismic, ect.) or hydrographic data. Camera systems are a relatively new addition to deep AUV systems. Currently, there are only a few companies, institutions, or government agencies that operate AUVs equipped with digital still cameras capable of survey to 1,000 meters or deeper.

/C-Surveyor-III/ AUV Being Launched (Courtesy of C & C Technologies, Inc.)

I am writing here primarily about C & C Technologies’ C-Surveyor AUVs, because I have the most access to these systems (a HUGIN 1,000, two 3,000, and a 4,500 meter systems).  Although the sensor payload of each of these AUVs may be slightly different, the basic payloads include an EM 2000 multibeam bathymetry system, Chip Edgtech subbottom profiler system, and duel frequency side scan sonar (120 kHz or 230 kHz dynamically focused and 410 kHz, or synthetic aperture). C & C’s has equipped three of these AUVs with digital still cameras (George 2009a).

In 2001, C & C began using the first commercial deep-water AUV in the Gulf of Mexico.  C & C surveyed the first of several shipwrecks with their AUV in January 2001 when the AUV passed the SS Robert E. Lee during a pipeline survey for BP and Shell. The SS Robert E. Lee was a passenger freighter sunk by the German submarine, U-166 during World War II. A continuation of the project led to the startling discovery of the U-166 in March of 2001. During the course of the survey two other historic shipwrecks, the Mica Wreck and the later designated Mardi Gras Wreck were imaged with sonar as well as four of SS Robert E. Lee’s lifeboats. Between January 2001 and January 2012, C & C collected over 246,000 line kilometers of deep-water AUV data, enough to circle the earth more than six times at the equator. These have included surveys of over 30 deep-water shipwrecks many of which are historically significant.

Photo Mosaic of /U-166/ Conning Tower and Deck Guns (Courtesy of C & C Technologies, Inc.)

Integration of digital still camera

In 2009, C & C began integrating digital cameras into their AUV fleet. The AUV photography system provides black and white still photographs of the seafloor while the vehicle travels at a speed of 3.7 knots. An image is taken approximately every 1.75 seconds which equates to one photo every 3.5 meters of travel at normal survey speeds (George 2009b). The length of the camera footprint is equal to 0.75 times the AUV with an aspect ratio of 4:3. The AUV is typically flown at 6 to 10 meters altitude during camera surveys with a typical tracklines spacing of 5 meter or less allowing for overlap of photos.

The first shipwreck imaged with the C & C AUV camera was the Ewing Banks Wreck in 2,000 feet of water. The near immediate success of the camera provided archaeologists with another tool to quickly assess and ground truth potential archaeological sites in deep-water. Soon other wrecks were imaged with the AUV camera including the Mardi Gras Wreck in 4,000 feet of water and the U-166, in 4,800 feet of water.

AUV Photo Mosaic of the Ewing Banks Wreck Draped Over Bathymetry (Courtesy of C & C Technologies, Inc.)

Advantages and Challenges

Three of advantages of the AUV camera system are a) the ability to take the collected images and efficiently mosaic the photos into larger geo-referenced images; b) the ability to combine those images with the other geophysical data to aid in interpretation and site analysis; and c) the ability to quickly ground truth targets detected with the geophysical sensors.

Several hundred photographs are collected during a typical camera survey and it is important to know what portion of the seafloor each photo represent. C & C developed a software application to sync the photos with the AUV navigation/positioning system and convert each photograph to a geo-referenced image. In addition, a post processing routine was developed to equalize the repetitive flash pattern produced on each photograph, adjust for spherical light spreading, linear attenuation, and flash scattering resulting from water column particulates. The result of this processing is nice evenly lighted geo-referenced images that can then be more easily mosaiced and imported into a GIS system.

Having the photo mosaic and geophysical data (e.g. side scan sonar, multibeam bathymetry, and subbottom profiler) collected simultaneously allows all the site data to be analyzed in conjunction. The photo mosaic can also be draped over the swath bathymetry to provide a three-dimensional photographic perspective of the site. Although individual photographs and ROV investigation may be required for detailed analyses of specific areas or features of a wreck site, being able to quickly see the bigger picture along with the geophysical data offers a larger perspective of a site for assessing site formation, artifact distribution, and other aspects of the site.

The AUV camera is also an excellent tool for ground truthing unidentified targets. Often potentially significant targets are detected with side scan sonar during an archaeological survey and a recommendation has to be made based solely on the geophysical data. Having the option to collect photos over select targets, helps remove most of the ambiguity in the interpretation.

Conclusion

AUV cameras are advantageous to both the survey industry and the advancement of deep-water marine archaeology.  Since the introduction of digital still camera systems into survey class AUVs, the technology has repeatedly proven its value, efficiency, and effectiveness.  Although the technology is still in its relative infancy, it has immediately demonstrated its benefit for deep water AUV surveys in ground truthing unidentified targets, inspecting previously known sites, and creating geo-referenced photo mosaics to analyze historic shipwreck sites.

What other potential archaeological uses or advantages are there for this type of technology?

References

  • George, A Robert
    • 2009a.  Sensor Upgrades for Deep-water Survey AUVs.  International Hydrographic and Seismic Search, (August): 32-33.
  • George, A Robert
    • 2009b.  Integrated High Resolution Geophysical and Photographic AUV System.  Oral presentation given at the IMCA Annual Seminar, Rio De Janeiro Copacabana, Brazil (November).

All images courtesy of C & C Technologies, Inc.

LiDAR: Pushing the bounds of a technology or using what we have effectively?

The literature surrounding the use of LiDAR, light detection and ranging, imagery can often be disjointed, vague, and impractical for its application in archaeological investigation.  Wanting to utilize the available data, I became frustrated with the lack of literature that described a basic methodological approach to using LiDAR.  The most common usage for LiDAR in archaeological contexts continues to be identification of sites and associated features.  Recent interest in LiDAR’s ability to aid in the monitoring of conditions on archaeological sites offers another opportunity to employ the available datasets (Challis et al. 2008).

LiDAR, light detection and ranging, is the constant transmittal of high-resolution laser light to the ground surface, with the time differential of each pulse recorded at the receiving station attached to a low-altitude aircraft (Fennell 2010:6-7).  The accuracy of the method varies dependent on location and how the data was gathered; essentially, a micro-topographic map of the bare surface of the site and surrounding lands can be produced for archaeological analysis.  LiDAR has been used in multiple case studies including both prehistoric and historic archaeological surveys with and without vegetation cover (Fennell 2010; Harmon et al. 2006; Petzold et al. 1999).

While the usage of LiDAR in archaeological contexts remains limited, the ways in which it is manipulated and more thoroughly realized continue to expand (Challis et al. 2008; Chase et al. 2011; Devereux et al. 2005; Devereux et al. 2008; Fennell 2010; Harmon et al. 2006; Rowlands and Sarris 2007).  The various techniques to extrapolate information include, among others, the application of hill-shading algorithms, the manipulation of illumination sources by direction and elevation, the alteration of contour intervals through arbitrary and relational settings, the creation of local relief models, the application of statistics in analysis to include nearest neighbor, quadrat, and chi-square, the variance of resolution between micro and macro glimpses of the landscape, and even the use of multiple color gradients (Challis et al. 2008; Chase et al. 2011; Devereux et al. 2005; Devereux et al. 2008; Fennell 2010; Harmon et al. 2006; Jaillet 2011; Rowlands and Sarris 2007).

Of course, where there is potential…there is also pitfall.  Some of the more common issues with LiDAR that deter it from a more widespread usage include the potential for data overload, inconsistency in its interpretive value, human error or unfamiliarity with LiDAR, present surface imagery’s inability to cope with temporal and/or cultural association, and resolution issues (Harmon et al. 2006; Jaillet 2011).  Another point worth noting is that while it is without doubt a useful tool in the archaeological toolbox, it continues to be a method that works best in conjunction with other archaeological methods to include other remote sensing techniques, historic documentation and field investigation (Fennell 2010; Harmon et al. 2006; Jaillet 2011; Kvamme et al. 2006).

At this point, we come to the crux of the matter: what are we doing with LiDAR?  In order to get at this question, we could go back to the algorithm.  The algorithm most commonly discussed in the literature of LiDAR deals with the language of computers and programming.  The meaning, in most instances, is in reference to the computer science behind its analysis and the GIS, geographic information systems, functions used to analyze it.  While a great deal has been learned and a great deal more will be learned using this standard definition, I would ask that we apply the most basic ideas behind mathematical induction and recursive relations to our methodological approach to LiDAR analysis.

One solution would be to apply a back-to-the-basics approach involving the basic recursive algorithm of Divide-and-Conquer.  Using the Divide and Conquer Algorithm, one would break the larger problem down into two more manageable questions.  What can we do with LiDAR, in addition to we have already done?  How do we go about doing it, in the most basic sense?  It is the second question that appears to be the one plaguing the archaeological community most, as we have excellent examples worldwide of what can be done with LiDAR and archaeologists are continuing to apply it in innovative ways.

We need to come to a consensus on the variables that we are trying to measure using the LiDAR dataset.  One way to go about this would be quantification of the variables using archaeological signatures that essentially typify features common to historic and prehistoric site types.

Essential to the idea of the Divide-and-Conquer algorithm is its parallelism, its ability to be used for multiple purposes, just as we know LiDAR can be.  The same set of variables can be combined in differing ways to represent the different archaeological signatures expected of different archaeological resources.  For example, a historic agricultural settlement might include linear features such as field lines, roadways, and waterways, as well as, polygon features such as structures and specific forms of vegetation.  A prehistoric quarry site might include polygon features such as borrow pits and distinctive topographic features advantageous to the process of quarrying for lithic resources.  The limits to the use of this technology are as of yet unmapped.

Essentially, what we need is a solution that is both mathematical and manual, a more efficient way to standardize LiDAR analysis.  One potential solution would be to compute a coding system to manage the variables and allow for the ability to analyze LiDAR datasets with reference to the individual and combined variables, which would, in turn, limit the number of possible outcomes to a manageable number that could be reviewed and manually analyzed by the archaeologist.

In closing, I ask the archaeological community to rethink the algorithm in LiDAR and continue to expand upon the ways in which we use this valuable tool.  Where to from here then… continue to push the bounds of this technology or begin to utilize what we have effectively?  Must we make this choice or can we begin to apply consistent methodological standards to our use of LiDAR, while pushing the bounds of possibility?

References Cited

  • Challis, Keith and Ziga Kokalj, Mark Kincey, Derek Moscrop, Andy J. Howard
    • 2008. “Airborne lidar and historic environment records.” In Antiquity. Vol. 82. 1055-1064.
  • Chase, Arlen F. and Diane Z. Chase, John F. Weishampel, Jason B. Drake, Ramesh L. Shrestha, K. Clint Slatton, Jaime J. Awe, William E. Carter
    • 2011. “Airborne LiDAR, archaeology and the ancient Maya landscape at Caracol, Belize.” In Journal of Archaeological Science. Vol. 38. 387-398.
  • Devereux, B.J. and G.S. Amable, P. Crow
    • 2008. “Visualisation of LiDAR terrain models for archaeological feature detection.”  In Antiquity. Vol. 82. 470-479.
  • Devereux, B.J. and G.S. Amable, P. Crow, A.D. Cliff
    • 2005. “The potential of airborne lidar for detection of archaeological features under woodland canopies.” In Antiquity. Vol. 79. 648-660.
  • Fennell, Christopher
    • 2010. “Archaeological Investigations and LiDAR Aerial Survey in Edgefield, South Carolina.” In African Diaspora Archaeology Network Newsletter.  December.
  • Harmon, James and Mark Leone, Stephen Prince, Marcia Snyder.
    • 2006. “LiDAR for Archaeological Landscape Analysis: A Case Study of Two Eighteenth-Century Maryland Plantation Sites.” In American Antiquity. Vol. 71(4).  649-670.
  • Hunter, William A.
    • 1960. Forts on the Pennsylvania Frontier (1753-1758).  Pennsylvania Historic and Museum Commission.
  • Jaillet, Angela S.
    • 2011. The People of Pandenarium: The Living Landscape of a Freed African American Settlement.”  Masters Thesis.  Indiana University of Pennsylvania.  Indiana, PA.
  • Kvamme, Kenneth L. and Jay K. Johnson, Bryan S. Haley.
    • 2006. Multiple Methods Surveys: Case Studies.  In Remote Sensing in Archaeology: An Explicitly North American Perspective.  Ed. by Jay K. Johnson.  251-268. University of Alabama Press.  Tuscaloosa, AL.
  • Rowlands, Aled and Apostolos Sarris
    • 2007. “Detection of exposed and subsurface archaeological remains using multi-sensor remote sensing.” In Journal of Archaeological Science. Vol. 34. 795-803.

3D Artifact Scanning @ VCU Archaeology

Virginia Commonwealth University (VCU) was awarded Department of Defense (DoD) Legacy funding for a three-dimensional (3D) artifact scanning project in 2011, which was developed in partnership with John Haynes, then archaeologist for Marine Corps Base Quantico (MCBQ).  The DoD Legacy program is designed to foster innovative approaches to the study, preservation, and stewardship of cultural remains—including archaeological objects—recovered on DoD facilities across the globe.

Our project involves 3D scanning of archaeological objects using a NextEngine Desktop 3D scanner in order to test and demonstrate the capabilities of this technology for its potential employment in ensuring DoD compliance with historic preservation laws.  Archaeological collections from DoD installations in Virginia, Maryland, and other regional repositories are the subject of the study. The Virtual Curation Unit for Recording Archaeological Materials Systematically (V.C.U.-R.A.M.S) consists of faculty member Dr. Bernard K. Means and several undergraduate students enrolled at VCU.

Virtual artifact curation has the potential for addressing a number of issues important to archaeologists. One issue is access to collections. The virtual curation project will enable researchers to access digital data files that allow full 3D observation and manipulation of an image and accurate measurement without requiring scholars to travel to a repository. Digital scanning of objects can save time for both researchers and for staff at curation facilities, while maximizing scholars’ access to collections.  Objects and entire collections that are now physically dispersed in more than one repository can be united through 3D digital scanning into a single virtual repository.

Visitors watch as Clinton King scans an artifact in the field at the Huntsberry Farm Civil War site outside Winchester, Virginia.

The NextEngine Desktop 3D scanner is designed to be portable and, as part of the Virtual Artifact Curation project, the potentials and capabilities of the scanner have been tested at several non-lab locations. We can go to places that are culturally and historically important to our country, scan objects at these locations, and make them accessible to a wider audience. We have been fortunate to scan archaeological materials from Virginia institutions such as Colonial Williamsburg, Jamestown Rediscovery, George Washington’s Ferry Farm, and Flowerdew Hundred, and at The State Museum of Pennsylvania in Harrisburg, Pennsylvania. Archaeological materials from these significant locations are certainly too fragile to be passed around among scholars and in classroom settings, but can be shared digitally.

With 3D scanning technology, important cultural items that belong to and must be returned to private landowners could be recorded and made available to scholars through virtual curation.  While owners of archaeological collections in private hands may not be willing to donate the physical objects located on their properties—perhaps identified through a compliance investigation—they may agree to “donate” the information inherent in their collections and make their items virtually accessible to a wider audience of scholars and others who might be interested. Virtual curation may also prove useful for cultural objects that are designated for eventual repatriation, if descendent groups agree to the scans of these items.

Courtney Bowles holds a bone tambour hook prior to scanning at George Washington’s Ferry Farm, Fredericksburg, Virginia.

Virtual curation of artifacts will prove critical for fragile objects by minimizing handling and “preserving” them digitally, especially when conservation funding is limited. Repeated digital scanning sessions can help conservators ascertain whether conservation treatments are working as intended—through highly accurate digital models taken of the same object at set intervals. This will enable the conservator to closely monitor whether there is continuing degradation of an object.

While digital scanning is an important tool for documenting the potential degradation of an object, the initial stages should precede any conservation treatments when possible. If an object is scanned prior to conservation treatments, a pretreatment scan of the object may be the “truest” image of the object that we will ever have. Conservation does not always produce an object, however stable, that represents its original state.

Sharing of data is certainly one of the strong points of the movement toward digital archaeological media. The ability to manipulate and move objects in three dimensions benefits researchers more greatly than static images ever can. Public and scholarly interaction with digital models can certainly foster a more reflexive archaeology. This would allow diverse observers to move virtual objects or travel through virtual worlds, creating a dialectical relationship between past and present—and, open interpretation and reflection up to a wider audience.

The scanning team in the Virtual Curation Laboratory at Virginia Commonwealth University, Richmond, Virginia. Left to Right: Clinton King, Bernard K. Means, Victoria Valentine, and Courtney Bowles.

Where do we go from here? How will 3D digital images of objects and artifacts alter people’s perceptions of what is “real” and what is “virtual”? This is something we plan to explore in greater detail in the coming months. Our project team maintains our own blog that regularly details and updates our progress with the scanning project: http://vcuarchaeology3d.wordpress.com.  Here, you can find more information about our successes and challenges with the virtual curation of artifacts from historic and prehistoric sites. We welcome your comments as well.