2026-11-11 –, Video Room
Introduction
Archaeological salvage projects in Mexico, defined by their direct association with public or private construction works that potentially affect cultural heritage, have become increasingly common in recent years as a result of the country's major infrastructure developments (Carballal Staedtier et al., 2000; López Wario, 2016). When such projects reach a large scale, they generate heterogeneous volumes of information, including registration tables, field logbooks, architectural drawings, and photogrammetric products, produced simultaneously by different technical teams and under the time constraints imposed by the construction schedule.In the Tren de Carga Archaeological Salvage Project, in southeastern Mexico, this structural condition calls for a system able to integrate, validate, and standardize excavation-phase information at a pace that manual review cannot sustain. This is especially critical given that the internal and external quality of the data underpins both its technical integration and the documentary evidence required to support institutional decision-making on the protection of archaeological heritage (Gupta, 2023). Such documentary evidence is, in fact, a project-closure requirement under INAH's regulatory framework, to be met through technical reports documenting the fieldwork carried out and the resulting archaeological record (INAH, 2024, art. 21).This work aims to document the geospatial architecture implemented by the geomatics liaison team of the Tren de Carga Project to automate the integration, standardization, and validation of archaeological information generated during excavation, and to examine how this architecture contributes to the institutional products (archaeological monument records, reports, mapping, and progress updates) that support heritage protection throughout construction.
Methodology
The architecture rests on four interrelated components. A relational database with spatial extension (PostgreSQL with PostGIS) consolidates all information in a single repository, queryable in real time by the different roles within the team (D'Urso and Aldrighettoni, 2024; González Tennant, 2016). QGIS serves as the environment for visualization, spatial verification, and cartographic production, including the semi-automated generation of institutional monument records through its Atlas tool.
Python and R are used to automate the processing, standardization, and validation of the inputs generated both in the field and in the office. Archaeological drawings digitized in CAD by the architecture team, including plans, profiles, and structural elements of the monuments, are processed through a Python script embedded in the QGIS toolbox. The script reprojects and corrects geometries, automatically identifies the corresponding monument, and classifies each entity according to a catalogue of more than 50 architectural types.
This process yields a set of standardized vector layers: floor-plan drawings, architectural projection lines, and excavation control grids, the latter including a field that records whether each grid cell was actually excavated. The generation and validation of monument records, together with the review of deliverables prior to integration, likewise draw on Python scripts and dedicated SQL queries that automate cross-checking and standardization tasks previously carried out manually.
In parallel, a set of R workflows built in R Markdown automate the ingestion, standardization, and integration into the spatial database of the project's main documentary inputs from the excavation phase: weekly field-crew logbooks, excavation registration tables, intervention forms, post-intervention process forms, and the vector layers derived from the digitized architectural drawings.
Each workflow follows a common sequence: authentication and download from cloud-based repositories; standardization of monument identifiers, dates, and column names; validation against catalogues of expected patterns, together with duplicate detection governed by explicit resolution rules (for instance, prioritizing the topographic source over the archaeologist's field record, or calculating a monument's area from its geometric shape using specific formulas, such as the ellipse formula for oval floor plans); generation of inconsistency reports; and final integration with the project's monument geometry layer through a spatial join.
This workflow automates the logical and archaeological consistency checks applied to the record without replacing human judgment in resolving discrepancies, and feeds directly into the generation of monument records and field construction-progress reports.
Results
The Tren de Carga Project's geospatial database currently holds records for more than 4,000 monuments addressed during the excavation phase, each of which generates multiple geospatial layers and associated products, such as registration tables, photographs, processed drawings, monument records, and progress reports. This multiplication of inputs per monument underscores the operational need for an automated system: at this volume, manually validating identifiers, dates, geometries, and typological fields for every monument would be unfeasible within the timeframes imposed by the construction schedule.
The automation implemented in Python and R allows inconsistency reports, institutional monument records generated through QGIS Atlas, and semi-automated progress reports to be produced systematically and with full traceability. These are the products that institutional bodies, namely INAH and DEFENSA, actually rely on to monitor the progress of interventions across each construction front and sector, and to support the issuance of Viable Construction Clearances (Vistos Buenos Viables de Obra), the condition that allows construction work to proceed without compromising the recording and protection of archaeological heritage.
Conclusions
This work illustrates how Geospatial Information Sciences function not as a mere technological add-on but as the methodological backbone that makes large-scale archaeological salvage, and the real-time institutional decision-making it demands, operable within the limited timeframes imposed by construction work (Siart et al., 2018).
The geospatial architecture presented here centers on the logical and archaeological validation performed on each layer prior to its integration, ensuring that the geographic information system does more than produce visually accurate maps and layers: it substantiates the archaeological content those layers represent with verifiable evidence (Menéndez Marsh et al., 2023).
Cristopher Enrique Foster Velázquez es Maestro en Ciencias de Información Geoespacial por CentroGeo y Licenciado en Ciencias de la Tierra por la Universidad Nacional Autónoma de México. Su trayectoria se ha enfocado en percepción remota, sistemas de información geográfica, programación y análisis de datos geoespaciales. Actualmente se desempeña como consultor geoespacial en SEDENA, dentro del Proyecto Tren de Carga, donde desarrolla soluciones para la gestión, procesamiento y análisis de información espacial, bases de datos y automatización de procesos. Su experiencia incluye además investigación oceanográfica, procesamiento de imágenes satelitales ópticas y SAR, y participación en campañas oceanográficas a bordo del B/O El Puma.