Evaluating the Robustness of Anti-UAV Detection under Controlled Fog Degradation: Fog-Aware Training and Clear-Sky Tradeoff

Name: Evaluating the Robustness of Anti-UAV Detection under Controlled Fog Degradation: Fog-Aware Training and Clear-Sky Tradeoff
Uploaded: 2026-10-04T11:19:26.711107+00:00
Duration: 328 s
Description: Researchers tested anti-UAV detection under ten levels of synthetic fog to see how performance degrades. They found that performance drops sharply, losing 96% of accuracy in thick fog. Training a model on both clear and foggy images improved robustness across all levels with only a small 9.1% drop in clear-sky accuracy, showing reliability is lost within a narrow visibility range.

Researchers tested anti-UAV detection under ten levels of synthetic fog to see how performance degrades. They found that performance drops sharply, losing 96% of accuracy in thick fog. Training a model on both clear and foggy images improved robustness across all levels with only a small 9.1% drop in clear-sky accuracy, showing reliability is lost within a narrow visibility range.

Open the full episode