FOSS4G IT & OSMit 2026

Given the expansion of the road network, wildlife-vehicle collisions (WVCs) are a growing global problem, representing a critical challenge for road safety and biodiversity. Traditional mitigation measures such as overpasses, underpasses, and fencing are often used, but their real effectiveness and high costs are still a cause of debate. With the advent of technological measures, new systems such as thermal cameras integrated with AI detection tools and alert signals are being tested. An experimental system that deploys this approach is currently being piloted in four municipalities of the Autonomous Province of Trento in Italy. While effective at identifying animals, vehicles, and humans' heat signatures through AI, this system still lacks wildlife species-specific accuracy. Within the TransWILD Biodiversa+ project, our study aimed to enhance the setup by adding camera traps to bridge this gap and provide data for comparison. Using QGIS as a decision-support tool, we identified optimal installation sites in two test areas. We integrated heterogeneous spatial data from different sources. During preliminary field surveys, the Gaia GPS app was employed to collect waypoints and tracks, which were recorded in the WGS84 (EPSG: 4326) reference system and exported in GPX format. Field data were then imported into QGIS and integrated with official cadastral datasets in SHP format and converted into the ETRS89/UTM zone 32N (EPSG: 25832) reference system. By overlaying these layers, a detailed map of the distribution of private properties in the candidate sites was created. The methodology followed a multi-step spatial process: 1) creation of an overlay map to identify public and private parcels, 2) proximity and buffer zone analysis to determine the optimal placement of camera traps and to maximize area coverage, and 3) data management and sharing to inform public authorities. The spatial outputs produced helped us in the decision-making process, demonstrating that FOSS4G is a powerful tool to communicate and create connections between technical teams and non-technical stakeholders working in public administration. Using FOSS4G was an advantage, as it allowed us to share information without licensing barriers. However, while preliminary tests with camera traps confirmed the high potential of the integrated monitoring system to prevent WVCs, we also encountered strong limitations. Privacy concerns regarding image capture from thermal cameras and complex bureaucratic procedures emerged as substantial obstacles, eventually limiting the collection of long-term research data and requiring the removal of camera traps. Despite these difficulties, the study emphasizes that interdisciplinary projects, which combine wildlife research with real-world social issues such as WVCs, depend heavily on the receptiveness and preparedness of public authorities. It is therefore essential to plan surveys well in advance to ensure compliance with regulations and local requirements. The FOSS4G approach and the decision to use tools such as QGIS are still the best choice for such collaborative initiatives, as they create an inclusive environment in which spatial data acts as a shared language for negotiation and transparency. Artificial intelligence systems designed to obscure people in recorded images could be implemented when using thermal imaging cameras, in addition to public awareness campaigns informing about the privacy safeguards and the benefits of these systems in terms of public safety. Future developments should integrate data ethics and administrative transparency into the project planning workflow to ensure a swift transition from technical design to implementation in the field, focusing on sustainability from legal and social perspectives.