Reference WWF Germany

With the AI-based monitoring solution “TrapScan”, we are supporting the nature and environmental conservation organization WWF Germany in analyzing wildlife camera data. TrapScan enables the efficient analysis of large amounts of video footage, the identification of animal species, and the clear visualization of observations. The solution helps systematically record species occurrence and movement patterns in the floodplain landscapes of the Middle Elbe and derive scientifically sound nature conservation measures.

Challenge

To observe wildlife, WWF Germany deploys numerous camera traps in the floodplain landscapes of the Middle Elbe. The resulting video data provides valuable information about animal species and their behavior, but at the same time requires a significant amount of effort to analyze manually. Since the cameras react to every movement, the recordings contain not only animals but also numerous irrelevant sequences, triggered for example by wind or vegetation movements.

For scientific analysis and nature conservation management, relevant observations must be reliably identified, documented, and evaluated. Manually reviewing the large volumes of data is time-consuming and ties up valuable human resources.

Implementation

Together with WWF Germany, we developed the AI-based monitoring solution “TrapScan.” The system automatically analyzes the video data recorded by wildlife cameras, detects animal movements, and classifies the animals captured using trained AI models.

We contributed our many years of expertise in audio and video analysis, visual animal recognition, and AI-based data analysis. Building on our experience in developing intelligent analysis and monitoring systems, we created a solution that is specifically tailored to the requirements of wildlife monitoring. Among other things, open-source models were technically optimized so that they run reliably in a production system.

The analysis results are made available in a web-based application and clearly visualized. Using interactive map and dashboard functions, observations can be evaluated, locations compared, and relevant video sequences viewed directly.

By automating the processing of large amounts of data, TrapScan enables significantly more efficient analysis of wildlife camera data and supports WWF Germany in data-driven monitoring of wildlife populations and habitats.

More information about the project

"TrapScan" was developed as part of the WWF projects along the Middle Elbe.