.png){kind=icon}
.png){kind=icon}
Return
Rabit – Rehabilitation of abandoned buildings and information tracking: a tool based on artificial intelligence
.png){kind=icon}
PRE
AI
G
A
DOI：10.1080/17452007.2026.2676851.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Many advantages are attributed to the rehabilitation of abandoned buildings, spanning environmental, social, and economic spheres. Official data indicate the presence of many uninhabited buildings in Brazilian metropolitan areas; however, information about these buildings is fragmented and dispersed. This work presents an Artificial Intelligence-based tool developed to collect and organize information on abandoned buildings available in various media, called RABIT (Rehabilitation of Abandoned Buildings and Information Tracking). The tool was developed in Python, combining semantic vectorization, neural networks, and Graph Theory to convert unstructured textual data into technical and territorial information. It consists of independent and integrable modules responsible for automated document collection, linguistic analysis, geolocation of information, cross-referencing with public databases, and filtering based on feasibility criteria. The case study on a specific building demonstrated the existence of multiple sources of information, which were scattered, fragmented, and available in various digital formats. The tool, developed using artificial intelligence, demonstrated its ability to track and organize information into previously defined categories, including information about the building, its location, and historical context. Given the complexity involved in rehabilitating abandoned buildings, the available information may be helpfull to all involved, especially to base feasibility study (investors), design development (designers) and construction (builders).
Keywords:
Abandoned buildings rehabilitation
information tracking
artificial intelligence tool
Journal
IF:
2.5
Papers:
183
Citations:
1.3K
Organization
Cited Papers