This study presents a novel application of artificial intelligence (AI) to detect whales in satellite imagery (SPACEWHALE), offering a new approach to assessing whale distribution and population trends in remote areas. Traditional monitoring methods are often constrained by accessibility, cost, and the challenge of surveying vast ocean areas. By applying advanced AI models for object detection, we use a semi-automated process to identify whales in high-resolution satellite imagery (e.g. WorldView-3 and Pléiades Neo), enabling large-scale and consistent monitoring of whale populations and movements. Human reviewers verify AI detections and assign species identification. The use of AI significantly increases detection rates and improves the efficiency of image analysis compared to human-only approaches.
We present an overview of SPACEWHALE applications across multiple regions, including New Zealand’s Southern Right Whale, demonstrating how it can help address knowledge gaps in remote areas. The method can also complement baseline surveys to support environmental impact assessments for marine development and conservation. We will discuss application of SPACEWHALE in Australia, where large, remote offshore areas and migratory whale populations present challenges for traditional survey methods.
Overall, AI-driven satellite monitoring offers scalable, repeatable surveys, supporting improved understanding of whale distribution.