GroundCount: Teaching Drones to Count Accurately with VLMs

A new framework, GroundCount, significantly improves Vision-Language Model (VLM) counting accuracy by integrating object detection, offering a practical solution for drones to precisely count objects in real-time.