1. Edit marks . Marks are placed in the data at regular distance intervals. This is to keep track of where the instrument is in space. When entered manually (as is sometimes necessary) marks are occasionally missed or extra marks are added. When a survey wheel is employed marks are still sometimes missed or erroneously placed. The marks need to be corrected so that they are consistent with distance.
2. Frequency high and low pass filters. When data are collected with these filters already applied, this step is skipped. However, many GPR raw data signals are recorded without frequency filters. If the high and low frequency “noise” that is outside of about 1 octave above and below the center frequency of the antenna used is not filtered out, the signal cannot be further processed. Typically these raw signals become skewed to one or the other side of zero with depth/time.
3. Gaining involves amplifying the signal with depth. It is necessary for raw data, and sometimes needed for data that was collected with low gains.
4. Frequency high and low pass filters . This often needs to be done repeatedly, especially if a raw signal was filtered and then gained (which amplifies some of the remaining high frequency noise).
5. Position correction . This entails finding the correct time position along each trace of the ground surface. This position varies as ground conditions change and the instrument drifts, and is best accomplished by adjusting each trace individually. This step is very time consuming and will be streamlined.
6. Background removal. This is a high-pass filter in the horizontal dimension.
7. Profile alignment. This entails reversing some or all of the profiles, if needed. It is sometimes necessary before slicing.
8. Velocity calculation . Hyperbola fitting is the most common method for calculating velocity for archaeological purposes. This entails locating hyperbolic reflections in profiles and measuring their width and height. This will be automated.
9. Stacking . This is simply re-sampling in a horizontal direction by averaging scans together.
10. Frequency filtering . This is often repeated to eliminate high frequency clutter.
11. Selective gaining . This is a subjective method that is often used to enhance archaeological features. Rather than equalizing the average amplitude, it is often useful to increase the gain only in selected areas where archaeological features are weak.
12. Hilbert Transform. This transformation is a technique for converting cosine waves to sine waves, effectively converting all the negative portions of reflections to positive and smoothing the transitional area. The result is one complete positive sine wave for every cosine wave (positive and negative wave couple).
13. Migration. This “migrates” the hyperbola tails and other reflections that are an artifact of the antenna’s wide angle of vision.
14. Slicing. This is the process of extracting values from a chosen depth interval and composing them into a 2D image or map of the subsurface on a horizontal plane.

Image Processing:

1. Mosaicking . This entails assembling the various blocks of data into one large, continuous image.
2. Destriping . This is the process of equalizing the means of each line of data to eliminate stripes. It can also be done by balancing the variances between lines.
3. Destaggering. This involves shifting every other row to correct for surveyor timing error, and filling in the gaps by interpolation. This process will be automated with the ability to shift odd, even, or other user-selected areas by distances defined in distance units.
4. Despiking . This is done by identifying data values over a specified threshold within a specified radius and replacing them with new values.
5. Edge-matching by means . This is done by matching the mean of one grid or block of data to the neighboring grid by differencing the means of the adjacent lines (of pixels).
6. Desloping . This is another way of edge-matching. Geophysical data often have drift and need to be corrected.
7. High pass filter . This is a convolution filter that subtracts the local average value from each pixel, thus enhancing high-frequency information.
8. Low pass filter . This is a convolution filter that smoothes the image.
9. Interpolation . There are a plethora of interpolation algorithms, many of which are not employed in COTS geophysical processing programs.
10. Sharpening . This is done by adding a high-pass-filtered version of the image back on top of the original, then dividing in half. It often enhances high frequency information, creating a sharper image.
11. Image registration . This step has typically been accomplished using a COTS GIS package. However, no single package offers the variety of transformations needed. For co-registration of geophysical images often only an affine transformation is needed, but for correct registration of an aerial photo or satellite image a projective or conical transformation is required.
12. Fusion. This step entails a variety of new methods for data fusion, including principal components analysis, color compositing and overlay, image classification, map algebra, Boolean logistical combinations, image segmentation, and object recognition. Note: These methods were pioneered in the SERDP Project (CS-1263) and will be one of the strongest, most innovative components of ArchaeoMapper.

The ArchaeoMapper software to be demonstrated in this project will address each of these steps and greatly streamline data processing and integration by allowing every step listed above to be completed within a single user-friendly software environment. In addition, ArchaeoMapper will implement processing routines that were identified and tested during the SERDP Project (CS-1263), but are currently either unavailable or inefficient in COTS software products used to process and integrate geophysical and other raster data. Graphical user interface-based “wizards” that automate many typical processing sequences will allow relatively inexperienced users to achieve professional-quality results.

The advancement of using a wide variety of remote sensing methods to identify and evaluate historic resources, including newly developed methods for processing and integration, has already been completed as part of the SERDP project CS-1263. Yet these technological innovations are not yet readily accessible to DoD personnel. In addition, while their value has been unambiguously validated by CS-1263, these results have not been demonstrated to DoD personnel or to the state, federal and tribal regulatory officials who, under 36CFR800, have the authority to review proposed DoD actions with respect to their compliance with applicable federal laws. In the demonstration and validation project, a group of DoD remote sensing practitioners will collect data at a carefully selected site and then process to integrate their results using ArchaeoMapper. Representatives of SHPO, ACHP, tribal groups, and other federal agencies will observe the process in the field and be informed (by means of concise, non-technical documents and presentations) of the project results.