Our experts launch the WISDAM app for researchers conducting wildlife imagery surveys with Dugong detection built into its AI
December 16, 2024
Earlier this month a new open source app was released that is designed for researchers conducting wildlife imagery surveys to produce data suitable for abundance estimation and spatial modelling from aerial photographic surveys.
Of interest to researchers involved in marine conservation is that their Marine Animal Detector AI has a large training dataset for nine animal classes. Whilst his does of course include dugongs, it can also identify whales, dolphins, turtles, sharks, rays, birds, fish and sea snakes.
This new tool for surveys to monitor species populations is called WISDAM. It was launched in Perth by Amanda Hodgson and Christophe Cleguer, both of whom were on the Technical Advisor team for The Dugong and Seagrass Conservation Project (DSCP).
The software is based on their extensive experience of developing imagery survey methods and knowledge of the data requirements for wildlife surveys.
A key aim is to enable non-experts to monitor species populations and their habitats. It will be useful to many projects conducted within DSCP. It allows users to import images from any aerial platform source. It then uses the power of georeferenced images to map image footprints and objects to real-world coordinates. Users can import detections generated by an AI process and WISDAM provides a user-friendly workflow to visualise, verify and enrich these objects with metadata. With customisable image and object labelling, WISDAM standardises the image review process and provides a range of data outputs, including AI training data. Quite simply, the more it gets used the more it improves.
Along with our experts, many others were involved including Dr Nat Kelly with the Australian Antarctic Division, and Dr Frederic Maire from the Queensland University of Technology. The software developer was Martin Wieser, a researcher and developer in the area of Photogrammetry and Remote Sensing.