OUR projects

Trapper keeper

With support from the WILDLABS Awards and Arm, the Open Science Conservation Fund with collaboration of BearID Project, Universidad de San Francisco de Quito and San Diego Zoo Wildlife Alliance  are launching Trapper Keeper, an open-source, AI-powered infrastructure for managing camera trap data at any scale — from remote, low-connectivity field sites to large research institutions. Building on our #TRAPPER 2.0 BETA, we aim to deliver a robust, scalable system with portable field servers and containerized deployments, enabling efficient biodiversity data management, processing, and sharing, and linking local monitoring efforts to global biodiversity networks.

Over the project’s course, we will:

  • Release a stable, open-source TRAPPER 2.0 for both small and large-scale projects.

  • Simplify deployment and enhance workflows, making the platform accessible to experts and citizen scientists alike.

  • Provide clear documentation and training materials to encourage adoption.

  • Develop rugged portable servers with AI capabilities (Raspberry Pi 5 + AI HAT+, NVIDIA Jetson).

  • Deploy energy-efficient Arm-based Ampere servers for larger installations.

  • Integrate advanced AI species classifiers and video tools from SDZWA and BearID.

  • Enable seamless publishing of standardized Camtrap DP datasets to GBIF.

  • Test and refine the platform with diverse global partners in Ecuador, Canada, Poland, and Venezuela.

A large-scale wildlife monitoring system using camera traps and artificial intelligence.

With support from NCBIR funds, TAXUS UL Sp. z o.o., the Museum and Institute of Zoology of the Polish Academy of Sciences, and the Open Science Conservation Fund will be working on the project
“A Large-Scale Wildlife Monitoring System Using Camera Traps and Artificial Intelligence.”

Project Summary:
The TRAPPER system, an open-source platform for camera trap data management, will be significantly expanded to support advanced wildlife monitoring and analysis. Key developments include enhancements to the data model and backend API to enable classification projects and video-based methods such as REM, distance sampling, and space-to-event. The citizen science frontend will be upgraded with new features like data upload, calibration tools for measuring distances in videos, and automated frame extraction at set intervals.

TRAPPER will integrate with the Trapper-TAXUS application via REST API to exchange data for AI-powered analysis of animal movement, distance, and abundance. Results will be stored at both object and video classification levels. Additionally, the data export module will be upgraded to support major ecological analysis methods and allow refined data queries. Exports will follow the global Camtrap DP standard for compatibility with analytical tools in Python and R.

TRAPPER AI: Scalable Citizen Science for Wildlife Monitoring in Sweden

A long-term scientific collaboration focuses on ecological research into the direct and indirect interactions between humans, large carnivores, ungulates, and vegetation. The partnership aims to advance research methodologies using camera traps, maintain and support the TRAPPER system, and further develop its capabilities.

As part of a collaboration between the Open Science Conservation Fund (OSCF) and the Swedish Hunters’ Association (Svenska Jägareförbundet) — under scientific guidance from researchers at SLU (Swedish University of Agricultural Sciences) — OSCF designed and implemented a new Citizen Science (CS) interface for the open-source TRAPPER platform. This extension is specifically tailored for non-expert users involved in organizing and processing camera trap data.

The new interface simplifies the process of organizing, sharing, and classifying data and transparently integrates AI functionalities such as object detection and species classification, making them intuitive to use. Its first implementation in Sweden marked the initial deployment of TRAPPER AI, intended to support a nationwide wildlife monitoring system. This system combines a distributed monitoring network (local hunter groups operating in their hunting areas) with a centralized data processing system (TRAPPER AI).

Improving the ecological connectivity in the Karkonosze National Park and its buffer zone

Through large-scale wildlife monitoring & AI-powered analysis, our joint project—”Improving the Ecological Connectivity in Karkonosze National Park and its Buffer Zone”—has provided data-driven insights into ecosystem dynamics and animal movement in Karkonosze.
Key Findings from Our AI-Powered Monitoring:

✔️ Smart Wildlife Detection – AI efficiently filters images and identifies 17 species, from foxes to wolves.

✔️ Tracking Seasonal Movements – Deer migrate to lower areas in winter, with wolves following their patterns.

✔️ Protecting Biodiversity – Connectivity models confirm the critical role of the buffer zone in ecosystem preservation.

We advocate for AI-driven conservation to become the gold standard in Poland’s national parks—enabling data-driven strategies to protect wildlife and sustain ecosystems for future generations

Microsoft AI for Earth

With support from Microsoft’s AI for Earth grant, researchers from the Mammal Research Institute of the Polish Academy of Sciences (IBS PAN) are transforming how wildlife is monitored and studied in Białowieża Forest—one of Europe’s last primeval lowland forests.

To tackle the growing challenge of processing massive volumes of data collected by camera traps, the team developed TRAPPER—an open-source, web-based system for managing, classifying, and sharing wildlife data. While TRAPPER streamlined many aspects of field research, manual data annotation remained time-consuming and labor-intensive.

The integration of artificial intelligence—particularly Microsoft’s MegaDetector—into TRAPPER now allows for automatic detection, filtering, and species classification in images and videos. This significantly increases research efficiency, reduces human workload, and enables large-scale, sensor-based ecological monitoring.

Beyond saving time and resources, AI opens new possibilities for scientific exploration. Researchers can now deploy a wide range of sensors in natural environments, confident that AI will help process and interpret the incoming data. With these tools, ecosystems like Białowieża can serve as living laboratories—not only for scientists but also for citizen scientists and conservation initiatives worldwide.