Scholarlink

<p>We evaluated the suitability of a commercial digital camera, the Nikon D5, for auroral observations. Using a monochromator, we measured the transmission characteristics of the RGB color filters and confirmed the sensitivity of the green and blue channels to the 557.7 nm and 427.8 nm emissions, respectively. Per-channel linearity and uncertainty were quantified using a calibrated integrating sphere as a standard light source, and software binning was shown to improve signal-to-noise ratio. After optical calibration, the camera was installed near Tromsø, Norway, to conduct actual observations. We implemented an automated imaging system driven by a deep-learning model that detects aurora in real time and activates the camera only during auroral occurrences, enabling high temporal resolution while greatly reducing data storage. Using 1456&#xa0;min of observation data, we quantitatively validated the accuracy of converting RGB counts to Rayleigh units by comparing with a co-located multi-wavelength photometer, showing that the Nikon D5 can estimate absolute brightness at the green and blue lines with uncertainties of a few hundred Rayleighs. These results validate an AI-assisted observing workflow and underscore the potential of commercial digital cameras for auroral research across citizen science and professional contexts.</p> Graphical Abstract <p></p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Calibration and observations of an AI-triggered commercial digital camera for auroral studies

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年5月29日

查看全文
  • Sota Nanjo,
  • Urban Brändström,
  • Satonori Nozawa,
  • Tetsuya Kawabata,
  • Keisuke Hosokawa

摘要

We evaluated the suitability of a commercial digital camera, the Nikon D5, for auroral observations. Using a monochromator, we measured the transmission characteristics of the RGB color filters and confirmed the sensitivity of the green and blue channels to the 557.7 nm and 427.8 nm emissions, respectively. Per-channel linearity and uncertainty were quantified using a calibrated integrating sphere as a standard light source, and software binning was shown to improve signal-to-noise ratio. After optical calibration, the camera was installed near Tromsø, Norway, to conduct actual observations. We implemented an automated imaging system driven by a deep-learning model that detects aurora in real time and activates the camera only during auroral occurrences, enabling high temporal resolution while greatly reducing data storage. Using 1456 min of observation data, we quantitatively validated the accuracy of converting RGB counts to Rayleigh units by comparing with a co-located multi-wavelength photometer, showing that the Nikon D5 can estimate absolute brightness at the green and blue lines with uncertainties of a few hundred Rayleighs. These results validate an AI-assisted observing workflow and underscore the potential of commercial digital cameras for auroral research across citizen science and professional contexts.

Graphical Abstract