Discovering and Recording Archaeological Features during a Survey Using an Unmanned Aerial Vehicle and QField: Application and Integration for Studying the Countryside Surrounding Perugia, Umbria (Italy) ^†

Abstract

One of the uses of Unmanned Aerial Vehicles (UAVs)—also known as drones—in archaeology is aerial reconnaissance, and they are usually used to detect and record the features of archaeological sites. This article focuses on the application and integration of drones with surveys supported using a mobile GIS (QField). The research results presented herein concern an area of the northeastern countryside of the city of Perugia, in the Upper Tiber Valley. Field walking was carried out in 2020/2021. The information contained in this article has been organized and elaborated in a GIS environment (Qgis) to produce archaeological cartography.

1. Introduction

The following contribution focuses on some of the results achieved thanks to the integration of new technologies in the traditional methodology of archaeological surveys [1,2] carried out in a sample area (Figure 1). The territorial context (50 square kilometers ca.) falls within the administrative boundaries of the Municipality of Perugia and is located northeast of the city. It lies along the Upper Tiber Valley, between the Umbrian Valley to the southeast and the Middle Tiber Valley to the southwest. The context can be said to be representative of the central Umbrian physical landscape, which is characterized by the alternation of narrow alluvial plains, gradually passing from hilly stratifications to low mountainous reliefs.

The aim of this research is to highlight the archaeological potential of a territory that has been little investigated. The role assumed by the Tiber River over the millennia is important in terms of settlement dynamics [3,4]. To this end, it was essential to produce a scientific contribution that could be integrated with the research carried out in the past, and that could be used by the local authorities for urban planning and the protection and enhancement of local historical and archaeological heritage through the computerization of archaeological data in a GIS environment.

2. Materials and Methods

For the acquisition of the archaeological data, the work was divided into several phases. First, a collection of all information relating to the historical and environmental context and the history of the studies in this area was undertaken; then, the entirety of this information was moved to a computer management environment to structure the database in QGIS to create a graphical and tabular information dataset (Information Reference used is Monte Mario Italy zone 2, EPSG: 3004). The GIS database system provided for the recovery of historical cartography [5,6], IGM orthophotos of the years 1954–1955 [7], technical and thematic cartography and WMS and WFS services for the use of orthophotos at 1:5000 and 1:10,000 scales. During this phase, the photointerpretation of discontinuities was carried out by comparing satellite and orthophoto images.

The construction of a solid database has made it possible to better structure survey activity. Datasets of information levels in the .gpkg format that were adopted during the survey were defined as follows:

• “Aree ricognizione” provides for the polygonal geometry layers of the funds both by surface reconnaissance (“Survey_finale”) and by drone (“Survey_drone”).
• “UT.gpkg” provides the polygonal geometry layers of the Topographic Units identified in the survey phase.
• “Ut_pt.gpkg” provides the layers with multipoint geometry, referring to the individual findings recovered in the survey in the years 2020 and 2021; each point corresponds to a finding collected or identified within the individual concentrations xzor in areas of holes [8] (p. 171). The information level provided a series of fields that were useful for identifying the single object collected, most of which were structured for value maps.

The Study Case

The mobile GIS used to collect the survey data is QField [9,10]. There are important advantages: the possibility of geo-localizing the individual evidence reduces the time in terms of costs as well as in the post-processing phase; the possibility of working in the field with the same properties of Qgis, both in terms of style and configuration of the project, without making the work heavy. By working with a common design base, it was possible to equip each individual researcher with their own electronic support, thanks to which the data entry could be recorded quickly.

Similarly, the value maps imported by Qgis have unique values and facilitate the compilation of fields without incurring terminological confusion or digitization errors. QField was used both for the collection of internal (“intra-site”) and external (“off-site”) materials concentrations [11] ([12], p. 33) and for the data related to areas where the collected fragments could be considered indicators of the population of the territory in ancient times, so-called “background noise” [13] (Figure 2).

The same procedure was followed for the acquisition of the remote sensing data using a UAV, for the 3D mapping of already known sites and for aerial reconnaissance. The latter, which is understood as a research activity aimed at identifying evidence and landscape patterns that are still unknown—not always detectable through surface reconnaissance—was designed to be one of the latest applications of Unmanned Aerial Vehicles in archaeology. Their use in archaeology is relatively new, but there are many scientific articles dealing with the topic and the different possible applications [14,15,16].

As for the traditional aerial reconnaissance [17], the identification of discontinuities with respect to the contemporary landscape by means of UAVs occurs in the form of negative or positive traces, generally in the form of visible traces in the ground [2], which depend on the type of crop, the rhythms of cultivation and the geomorphological characteristics of the soil, such as crop marks [18,19]. As for the activities carried out in the context under consideration, aerial shots were taken to produce 3D models, and the collection of oblique photographs was carried out. Each model is provided with a data sheet that defines the individual characteristics. The equipment used [20] is the Mavic 2 Pro equipped with a high-resolution sensor 1” CMOS, 35 mm equivalent 24–48 mm (f/2.8–f/3.8) with a resolution of 20 Mp (≤2 cm) [21]. The digital images were taken with an overlap of 80% and a sidelap of 60% with an angle of 90%. Then, the images were tested in the laboratory using the range imaging technique of photogrammetry Structure for Motion using Agisoft Metashape Pro software (v. 1.8.0). Due to the dense point cloud classification process, it was possible to create several Digital Terrain Models (DTMs). The creation of DTMs allows you to generate true-level curves that can be attributed to the ground. All these products have been included in the GIS system as the basis for some pre- and post-reconstruction considerations.

3. Results

The integration of the two applied methodologies used in this work, drone reconnaissance and a traditional survey, has produced interesting results. In addition, these applications have confirmed how even greater joint use of different methodologies can be the only correct and fruitful way to follow the archaeological landscape.

Four case studies highlighted by the integration of UAV and traditional surveys with QField are highlighted below for the mapping of known contexts and for aerial surveys.

The Case Studies

The first case study (Figure 3) regards the 3D mapping of a funerary monument with a square plan in opus caementicium covered with sandstone blocks and bordered by a fence in opus reticolatum (I BC–I AD) [22] (p. 395). The relief generated an orthophoto from which it is possible to notice a change in tone and color given by the different growth of vegetation.

The anomaly is also confirmed by the digital elevation model that highlighted a quadrangular structure and a smaller anomaly in the center. At the same time, a surface survey was carried out, but no material was collected.

The second case study concerns an example of how UAV aerial reconnaissance has allowed for the identification of very complex anomalies for buried structures, made visible in the form of “yellow on yellow” crop marks (Figure 4) [15] (p. 180). The drone survey allowed us to acquire oblique photographs and create three-dimensional models, giving back the orthophoto and the DTM. The photointerpretation of the anomalies showed a complex internal structure, not better identified via reconnaissance. In fact, the survey did not collect ancient material, probably due to recent alluvial sediments, called younger fill [1] (p. 155), which increase the soil level via progressive deposition, obliterating any archaeological deposits below a depth of 60/100 cm.

Unlike the previous case studies, the third example shows how the integration of survey and aerial reconnaissance using UAV has introduced significant data (Figure 5). In fact, after the aerial reconnaissance with UAV identified a series of perpendicular and orthogonal alignments, the survey carried out with QField allowed us to collect a lot of materials. A first observation in the field, followed by the laboratory study of the material collected and the analysis of its spatial distribution in the GIS environment, led to a hypothesis concerning the presence of an articulated buried structure, probably relevant to a residential complex (I sec. BC–IV/V sec. AC).

Finally, the last case study that we intend to present is also an example of how it is not always possible to use remote sensing systems. In this case, the UAV survey was not possible due to the presence of a no-fly zone. However, the criticality of the area did not prevent traditional reconnaissance assisted by QField. The collection of the materials carried out by sampling and its study in the laboratory enabled the identification of an important and complex buried structure pertinent to a rustic villa (mid I sec. BC–IV/V sec. AC). The concentration has a quadrangular shape, which is quite clear and has the dimensions of about 120 × 150 m (13,800 sqm, that is 1.4 ha = 1 × 1 actus). The analysis of the distribution of the materials in the GIS environment has allowed us to hypothesize the internal subdivision between rustic pars and Dominica pars (Figure 6).

4. Conclusions

Even if it is preliminary research [22], the results have made it possible to reach the main objective, namely, to highlight the archaeological potential of the area. There are numerous settlement contexts identified in the diachrony: from the prehistoric period, with frequent areas mostly corresponding to alluvial terraces, to the medieval one. The greater yield, as is known, is the one resulting from the Roman period, from the Late Republic to the Empire and from Late Antiquity. In general, the identified concentrations are 32.

The results have made it possible to add much more information and data relating to the Carta Archeologica dell’Umbria (CAU) [23], which has been included in the Geoportale Nazionale Archeologia (GNA). This new information has also expanded the existing data [24,25].

The results, in addition to providing a clear example of how the good integration of multiple methodologies contributes to the contribution of significant results, offer a fundamental contribution to knowledge concerning ancient population dynamics.