VEMOP has generated large amounts of different data originating from the surveys with both the single- and multi-channel GPR devices; from the soil and sediment analyses in the laboratory and of course from the monitoring sensors installed underground at our four test sites.
As previously reported here (link), Erich has developed the custom-made Python software Schlitzi+ to organize, visualize, and analyze this information. During the last months, he expanded and refined Schlitzi+ further, adding ways to quickly and easily estimate EM signal velocity and to integrate the sensor data from our monitoring stations as well as the field diary and fieldwork pictures. This provides us with the basis required for the actual data interpretation.
One of the most important elements when processing GPR data is to know how fast or slow the EM signal travels underground. Variations in speed largely correspond to variations in water content and this is one of the reasons why we are so interested in it. In Schlitzi+, Erich has implemented a function to estimate the EM signal velocity using a method called hyperbola fitting – that is utilizing the geometry of reflection hyperbolas to determine the signal velocity. While hyperbola fitting is certainly nothing new, Schlitzi+ allows us to quickly scroll through complete datasets and analyze and store velocity values as specific points, essentially generating models that can then be used for the processing of the GPR data. The plotted models can also help explain some of the variations we see in the quality of the data sets.
Schlitzi+ will also help in integrating the in-situ monitoring data we collected during the GPR surveys. If you remember, each of the four monitoring sites had four to six soil sensors installed in the ground, as well as rain gauges set up above the surface. These sensors were set to measure volumetric water content, bulk electrical conductivity, temperature, and precipitation every 30 minutes. During a total of 16 months, this led to tens of thousands of data points, stored in long excel tables that are rather difficult to handle. Erich’s additions in Schlitzi+ allows to not only access these data and interactively plot them, but they can also be compared with other data sets collected under different conditions – with just a few clicks. This greatly speeds up the data analysis, in contrast to having to manually select and plot all these sensor data in excel one by one.
Another helpful addition comes in the form of a button called ‘get survey information’. This function was created to display both images and videos taken during fieldwork as well as written information on relevant observations recorded in a field diary.
Working on Schlitzi+ during the last two years, Erich has mainly focused on providing custom solutions to VEMOP. This little program, however, contains a wide range of useful functions that could (and should) be integrated into a standard workflow when working with GPR surveys in archaeology. To further improve Schlitzi+ and provide a more general, reliable and user-friendly version ready to be used by a wider community, Erich will continue to work on cleaning and documenting the code in a separate project. And we are very pleased by this somewhat unexpected, yet very important contribution of VEMOP.