In-field real-time data processing is still today a crucial factor that could definitively boost the success\ud
and spread of shallow seismic reflection and multichannel ground-penetrating radar (GPR)\ud
methods in the near-surface geoscience community, as it allows efficient data acquisition and cost-effective\ud
results, especially for modern surveys generating large volumes of data in a short time. To\ud
help fulfil this need, we present a cloud-computing solution combining the powerful computational\ud
capabilities of a cloud infrastructure with a subsurface imaging workflow based on a parallelized\ud
grid version of the Common-Reflection-Surface (CRS) stack, a macro-velocity model independent\ud
imaging method that is very suitable for real-time imaging, as its data-driven implementation avoids\ud
time-consuming human interaction in prestack velocity analysis. Our portal is accessible from the\ud
field by any mobile computer having a wireless data connection. The user-friendly web-browser\ud
interface allows, already during acquisition, uploading of the recorded data to remote computing\ud
facilities, where a quality control (QC) data analysis report is automatically produced. When the\ud
number of uploaded shot records is large enough to produce a subsurface image, stacking, velocity\ud
model determination and prestack time migration can be performed fast and highly automated,\ud
optionally after applying some preprocessing (e.g., gain, trace balancing and filtering) using the\ud
preprocessing toolbox of the portal. To demonstrate the use of the presented system, we simulate\ud
in-field data processing for an already published shallow seismic data set and compare our results\ud
with the original ones. The data set, consisting in seismic P-wave data, was collected to image\ud
Palaeozoic bedrock at the Flumendosa River Delta, Sardinia (Italy), with a data acquisition set-up\ud
driven by the expected bedrock depth, which, turning out wrong, prevented a detailed velocity\ud
analysis and a good time migration. Using the QC data analysis reports to run the imaging routines\ud
of our cloud portal we produced within less than an half an hour stacked and migrated sections very\ud
close to those published in the original work. These results, if obtained in the field, would have\ud
allowed an immediate update of the experimental set-up. Therefore, we are optimistic that the proposed\ud
cloud-computing solution (or similar systems) can boost the spread of shallow seismic reflection\ud
and multi-offset GPR surveys in near-surface investigations, similar to what happened in the\ud
field of electrical and electromagnetic surveying after suitable real-time imaging and data analysis\ud
systems emerged