• Geomatics at U of A
• University of Arkansas Spatial Library
• New sources for NW Arkansas Imagery
• Animation at U of A

Home > Streamlined Archaeo-Geophysical Data Processing and Integration for DoD Field Use > Implementation Plan

Abstract | Problem Statement | Technical Description | Implementation Plan | References

Each of the sensors now in use in archaeological studies has a specific software package that has been developed, often by the sensor designer, to process that instrument’s data. Each of these packages has a unique (often idiosyncratic) data structure, user interface and workflow requiring that a practitioner master each software package. Most of the data structures produced by the instruments and software are not compatible, requiring transformations that can lead to data loss. None of the software is designed to permit data fusion. ArchaeoMapper will bring together the most useful functions of the current ad hoc approach and new processing and integration techniques that have already been demonstrated in SERDP project CS-1263. The algorithms to be included in ArchaeoMapper have already been tested by SERDP project participants (albeit in the context of the type of manual, highly iterative processing described above). ArchaeoMapper will be assembled using the MATLAB and Simulink development environment. MATLAB is a widely-adopted, high-level (compared to the low-level C language) technical computing language and interactive environment for algorithm development, data visualization (including interactive 3D volumetric display), data analysis, and numerical computation. It is considered a standard for technical computing and has been used by many federal agencies and labs – including Argonne National Laboratory, Los Alamos National Laboratory, National Aeronautics and Space Administration, the U.S. Air Force, and the U.S. Navy – for both rapid prototyping of complicated processes and for application deployment. MATLAB and its accompanying toolboxes provide extensive suites of relevant, tested and commercially validated algorithms for signal and image processing, optimization, curve fitting, interpolation, and more. For example, the Signal Processing Toolbox provides interactive and programmatic tools for designing, visualizing, analyzing, and applying the complex filters required to “clean” GPR data. The volumetric data cubes generated by GPR sensors are easily stored and indexed using n-dimensional tensor notation and may be visualized in interactive graphic displays that provide programmatic tools to extract and view cross-sectional planes. The Image Processing Toolbox provides the tools necessary to geometrically correct (using, for example, conformal, affine, projective or user-defined transforms) and resample the various image data sets required in ArchaeoMapper. Two-dimension frequency filter design tools in the toolbox facilitate rapid and efficient noise removal from images. Developing these capabilities in the C or Java language would either require designing each of them from scratch or calling on libraries from a variety of sources. Either prospect would significantly increase the cost of building a complex application like ArchaeoMapper. However, these capabilities, combined with custom Graphical User Interface (GUI) design utilities that are part of the MATLAB programming language make MATLAB a very effective platform for prototyping sophisticated applications such as ArchaeoMapper. Field deployment of ArchaeoMapper to computers that do not have MATLAB installed is enabled by the MATLAB Compiler which is able to create a stand-alone, deployable .exe file.

The algorithms that will be assembled in MATLAB are summarized below.

Algorithms to be coded in MATLAB

Integration of Algorithms

Any application developed with the MATLAB tools will run quickly and effectively within the MATLAB environment. It is also possible with the MATLAB Compiler to assemble the entire program, including all graphical components, in the C/C++ language and deploy it as a stand-alone application running independently of MATLAB in Windows, Linux, or UNIX operating systems without a MATLAB license. The figure below shows an example of a stand-alone pattern detection tool built in MATLAB.

Pattern analysis GUI built in MATLAB
(click on image to see larger file)

Simulink is fully integrated with MATLAB and allows the user to quickly create, model, and maintain a detailed block diagram of a system using a comprehensive set of predefined blocks or blocks defined by MATLAB code. Simulink will be used in the application design and debug stages to interactively step through the process, analyze the results, and modify the algorithms as needed. It could also be used by experienced analysts as a more flexible and extendable interface.

Simulink block processing model and associated MATLAB code
(click on image to see larger file)

Finally, MATLAB provides a wide range of commands to export data to a very large number of formats from user-defined text files to common GIS formats. These commands, along with export capabilities specifically designed for the processes developed in CS-1263, will allow practitioners to exchange data with other sensor-specific applications.

ArchaeoMapper will provide future practitioners a common, consistent user interface so that it will not be necessary to learn multiple systems and interface idiosyncrasies. The next step is to demonstrate these new methods to a broader audience of practitioners and regulators. This ESTCP project will demonstrate and validate our newly developed data processing and integration methodologies to DoD and civilian agency personnel in a field setting and will also involve SHPO, THPO, ACHP historic preservation regulatory office personnel.