K. Palágyi Sylvia szerk.: Balácai Közlemények 2005/9. (Veszprém, 2005)

NEDUCZA, BORI SZLÁV - PATTANTYÚS-Á., MIKLÓS: Combined radar mapping of North-East corner of the roman Villa-estate, Baláca

In the next year, only the GPR method was applied. We carried out test measurements by profiling in two perpendicular directions in order to check whether double measuring time and effort give double amount of information. This trial was unsuccessful because the site we chose was rather disturbed by earlier agricultural activity and the remnants were mainly destroyed. The interpretation of the two-directional profiling did not correlate well, no walls and buildings could be localised, except of one part of the socalled round-wall (stone-fence).

5 In the year 2001, the dense GPR profiling was going on in the same way but only in one direction (with some perpendicular test lines), using the same instrument. The dimension of the survey site was 2500 m

2 with line separation of 1 metre. The location of measuring grids are shown in Fig. 1. The interpretation of sections has been executed one after the other by hand, because the collected data of the Noggin instrument cannot be processed later by advanced software on PC (no digital data storage). This is a disadvantage of this instrument, although the fieldwork is fast and results can be interpreted immediately on site. For this reason, we made some trial in this year with the pulseEKKO system that has a digital data storage ability and the measured raw data can be later gained, filtered, enhanced and processed. The frequency we used was 450 MHz. This test was successful, therefore in the next year we applied this equipment for radar mapping, but using also the Noggin for comparison reason. Survey in 2002-2004 - The NE corner

6 7 With the Noggin-survey, we moved in the NE direction about 90 m apart from the actual excavation site, that is from the line of the main building. Starting from this point, we measured grids of 50 x 50 m with a line spacing of 1 m using the pulseEKKO1000 type instrument with 450 MHz antennae. This higher frequency means usually a bit less penetration depth but much better resolution. The fieldwork is going on by moving the antenna along the line there and back, while data acquisition is continuous. The distance intervals are marked by hand. The sampling interval along the line -depends on the speed of the antenna movement, so the speed should be uniform and slow enough that the interval could be about 5-10 cm. The speed of data storage depends on the capacity of the PC. Usually only the quality of data can be controlled on site. For the interpretation some processing procedure should be executed on the sections. Once the data are stored, they should be transferred to an indoor PC for making some preliminary modification. The direction of each second section has to be reversed then time-zero correction and low-frequency (DC) filtering has to be performed. The positioning of each section has to be corrected according to the markers. These are the preparatory steps. Radar sections are usually checked individually before general processing. The processing consists of signal gain, frequency filtering and sometimes deconvolution.