Despite rapid advances in drone technology and multirobot coordination algorithms, few systems have been validated in real-world wildlife monitoring scenarios. This study presents a framework for autonomous collection of high-quality, multi-perspective data on gregarious animals in their natural habitat. Our approach is based on a particle swarm optimisation algorithm that computes the positions of drones according to the locations and orientations of the animals, ensuring effective non-intrusive observation for biological data collection. The system was deployed during a field campaign at Ol Pejeta Conservancy, Kenya, with 12 missions using three commercial off-the-shelf drone platforms. The data collected confirms that a drone swarm can effectively capture multi-perspective imagery of zebra herds to support wildlife conservation efforts. However, the computing time of our particle swarm optimisation algorithm reduced the quality of the monitoring, highlighting the need for a more responsive system for our next field campaign in mid-2026. Additionally, the experience of deploying drone swarms in the field offers valuable insights for future deployments and system improvements, particularly when operating in harsh and unstructured environments.