Input-Adaptive Gating of a Dehazing Front-End for On-Device Perception in Smoke-Obscured Environments

Two-stage vision pipelines often place an enhancement network before a task network, on the assumption that a cleaner input produces a better output. We evaluate this in a firefighter assistance pipeline, where a dehazer precedes an edge detector that renders smoke-filled rooms as structural outlines. Both were designed for a Raspberry Pi 4, at 355K and 23K parameters, and quantized to UINT8 via TensorFlow Lite. The float dehazer reaches 18.60 dB peak signal-to-noise ratio (PSNR) on held-out real smoke against 13.60 dB unprocessed and 17.08 dB for an AOD-Net trained on the same data, and the edge detector reaches an F-measure at optimal dataset scale (ODS) of 0.738, outperforming an optimized Canny's result of 0.692. Dehazing improves edge extraction under dense smoke but degrades it on clear and lightly hazed frames, where the dehazer discards more detail than the haze obscures. We therefore run the dehazer only when a dark channel haze estimate exceeds a threshold, a 10.1 ms test that lets the pipeline save 469.6 ms on the dehazing stage. With the four haze levels weighted equally, float32 gating is more accurate than either fixed decision, at 0.675 mean ODS against 0.664 for always dehazing and 0.630 for never dehazing. With the deployed uint8 models, the same gate reaches 0.601 mean ODS, versus 0.549 for always dehazing and 0.607 for never dehazing. It reduces the estimated equal-weight mean per-frame time on the Raspberry Pi from 569 ms to 321 ms, and on clear frames increases the frame rate fivefold, from 1.8 to 9 frames per second, but under dense smoke—the condition this application primarily targets—the gate takes 570 ms, saving no time.