Fig. 4: Detection, Tracking, and Classification of Moving Targets in Sparse Forest.
From: An autonomous drone swarm for detecting and tracking anomalies among dense vegetation
Under simple conditions (no or little occlusion), the swarm detects and tracks the target that appears most abnormal (a moving vehicle). a A satellite image with the ground-truth path of the vehicle (yellow) and the path of the tracking swarm (swarm’s center of gravity, blue line), individual drone positions of the swarm at time t (small light blue circles), best sampling position at time t (small yellow circle), and drone movement between time steps t − 1 and t (white arrows). b, c Close-ups at various times (drones encircled). d Visual results of RGB and anomaly integrals, blobs detected, and classification results at various waypoints (indicated by circles in a and by matching frame colors in d). This experiment was recorded in Supplementary Movie 2(Experiment I).
