A Real-Time Mobile Application for Cattle Tracking using Video Captured from a Drone

In many countries, governments have issued laws on cattle traceability, which require farmers to keep track of their livestock. While there are many algorithms to monitor cattle in confined indoor environments, monitoring cattle of a large number in outdoor pastures is still an open research problem. In recent years, with the advancement of unmanned aerial vehicles, e.g., drones, cattle monitoring can be based on images captured by drones. However, the challenges include image analysis in real-time, and keeping track of a dynamic scene (cattle movements) based on a moving sensor device. In this work, we develop an iOS application that can send captured herd images to our deep-learning-based server for cow segmentation and counting. Furthermore, the app can guide the operator to control the drone flight route and viewing perspectives.