Raw Imagery Impacting Your AI: Should You Care?
Read the original on arXiv AI →The paper investigates how raw or minimally processed satellite imagery affects onboard AI object detection for space missions. By systematically degrading Very High Resolution Maxar images in terms of Signal‑to‑Noise Ratio, Modulation Transfer Function, and Ground Sampling Distance, the authors evaluate three lightweight detectors—YOLOv5s, YOLOX‑S, and NanoDet—on the resulting data. Results show that image quality impacts detection performance in a degradation‑specific way, with GSD consistently shifting performance, while MTF and SNR effects vary by model and resolution; severe blur‑plus‑noise combinations cause the greatest losses.
Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv AI.