Thermal anti-UAV detection improves with balanced memory replay
Catastrophic Forgetting in Sequential Thermal Anti-UAV Detection: The Role of Scale-Conditioned Gradient Imbalance
Computer Vision and Pattern Recognition
Summary
Detecting drones with thermal cameras needs constant updates to handle new data. The authors studied how retraining a drone detector causes it to forget old detection skills, especially for different drone sizes. They found that forgetting mainly happens for large drones because training updates focus unevenly on small targets. Using a balanced memory of past examples sorted by drone size helps reduce this forgetting, keeping detection of large drones more reliable. Their results suggest managing training examples by drone size can improve continual learning for drone detection systems.
What this means in practice
- •For drone defense developers: Maintain effectiveness of thermal drone detectors over time by replaying past examples stratified by drone size to reduce forgetting large targets.
- •For security system engineers: Improve robustness of UAV detection systems against changing environments by using scale-aware memory buffers during detector updates.
Authors
Khac Duc Giang Nguyen, Seyed Sahand Mohammadi Ziabari, Ali Mohammed Mansoor Alsahag
Abstract
Counter-UAV systems based on thermal infrared detection must stay accurate as operational datasets evolve, yet sequential fine-tuning causes catastrophic forgetting of prior tasks, a problem that remains insufficiently characterized in this domain. This continual-learning study measures the stability-plasticity trade-off in YOLOMG, a YOLOv5-based detector run as a single thermal-infrared stream with the motion channel disabled, trained sequentially across three anti-UAV benchmarks of rising scale difficulty: Anti-UAV-RGBT, Anti-UAV410, and CST Anti-UAV. Naive fine-tuning on CST yields a Forgetting Measure of -0.605 against the Stage 1 ceiling, corresponding to a 90% capability loss, with -0.572 occurring in Stage 3 alone. In contrast, knowledge distillation from a frozen teacher is associated with FM = -0.033 +/- 0.004 across three seeds, corresponding to 95% retention. Because no Stage 2 no-KD control is included, this result establishes retention under KD training rather than a causal KD effect. Per-stratum analysis shows large-target detection collapsing to near zero within the first epoch, despite an inter-stage cosine similarity of 0.987 over the gradient-updated weights, pointing to scale-conditioned gradient imbalance, rather than weight drift, as a candidate mechanism. Scale-Stratified Herding (SSH), a 300-exemplar buffer balanced across four UAV size strata, roughly halves the forgetting (FM = -0.605 to -0.311) and keeps large-target detection non-zero. An ablation attributes the gain primarily to scale stratification rather than herding: random-stratified replay performs at least as well (FM = -0.221 versus -0.311 for SSH). These replay results are single-seed and should therefore be treated as preliminary.