Get Started
Home
Topics
Search
Library
Research questionHow can systems count and track people in dense crowds from moving-drone video under changing views and lighting?Drone motion changes viewpoints and spatial relationships across frames, while illumination and complex crowd motion make it difficult to maintain identities and accurate counts. Dense scenes also increase the risk that errors in one person’s identity or trajectory affect others.
AI
Computer Vision
Evaluation & Benchmarks
Image & Video Processing
Machine Learning
Research Paper
Technology
Latest papersRecent research connected to this question, newest first.Video Individual Counting and Tracking from Moving Drones: A Benchmark and MethodsThe source provides evidence from MovingDroneCrowd++, a large-scale moving-drone video dataset spanning flight altitudes, camera angles, illumination conditions, and dense crowd motion, together with reported counting and tracking experiments. Its conclusions are limited to the supplied benchmark and evaluated methods; the dataset, code, and pretrained models are reported as available.research paper · Sep 4, 2026
Related questions
How can individual animals be counted across sparse camera-trap frames when tracking labels are scarce?How can multiple aerial robots identify and track individual wildlife using onboard RGB cameras under decentralized, low-bandwidth coordination?How can robots navigate dense crowds when pedestrians respond differently to their presence?How can language-guided multi-object tracking preserve identities and context across long-horizon actions in 360° video?