[Paper Review] All Sky Camera instrument for night sky monitoring

Dušan Mandát, M. Pech+6 more|arXiv (Cornell University)|Feb 19, 2014

Impact of Light on Environment and Health1 references11 citations

TL;DR

This paper presents the All Sky Camera (ASC), a low-cost, autonomous instrument for monitoring night sky brightness and cloud cover using a fish-eye lens, CCD camera, and automated data processing. It achieves accurate cloud fraction (±2% for <20%, ±5% for >80%) and sky brightness (in mag/arcsec²) by calibrating star positions and analyzing star visibility across the sky, enabling long-term, remote monitoring of astronomical site quality.

ABSTRACT

The All Sky Camera (ASC) was developed as an universal device for a monitoring of the night sky quality and night sky background measurement. ASC system consists of an astronomical CCD camera, a fish eye lens, a control computer and associated electronics. The measurement is carried out during astronomical twilight. The analysis results are the cloud fraction (the percentage of the sky covered by clouds), night sky brightness (in mag/arcsec2) and light background in the field of view of the camera. The analysis of the cloud fraction is based on the astrometry (comparison to catalogue positions) of the observed stars.

Motivation & Objective

• To develop a low-cost, autonomous system for continuous monitoring of night sky conditions at remote astronomical sites.
• To address the need for passive, non-interfering instruments that do not affect high-energy cosmic ray detector operations.
• To enable reliable measurement of cloud fraction and night sky brightness in real-world conditions, especially at candidate observatory sites.
• To ensure long-term operation with minimal maintenance using weatherproofed, solar-powered electronics.

Proposed method

• The ASC uses a 1600×1200 pixel CCD camera (ICX274 AL) with a 185° fish-eye lens and 16-bit ADC to capture full-sky images during astronomical twilight.
• A calibration procedure maps pixel positions to celestial coordinates using a movable light source and polynomial fitting, validated with star catalog positions.
• Dark frames are subtracted from sky images to reduce noise, and zenith-angle-dependent sensitivity corrections account for vignetting, distortion, and atmospheric extinction.
• Cloud cover is calculated by comparing detected stars to cataloged stars (Yale Bright Star Catalog BSC5) within 1° angular tolerance across 70 sky segments.
• Night sky brightness is calibrated against a Unihedron SQM-LE meter, enabling conversion of pixel intensity to mag/arcsec² in the visible spectrum.
• System autonomy is ensured via a miniPC with USB I/O for iris and power control, internal heating for winter operation, and solar or grid power.

Experimental results

Research questions

• RQ1 How accurately can an all-sky camera estimate cloud fraction using star visibility and catalog matching?
• RQ2 What is the impact of lens distortion, vignetting, and atmospheric extinction on night sky brightness measurements?
• RQ3 To what extent does moonlight or weather conditions like fog, rain, or snow affect the reliability of ASC measurements?
• RQ4 Can the ASC system operate autonomously for months without maintenance in remote, harsh environments?
• RQ5 How does the algorithm's uncertainty vary across different cloud cover levels?

Key findings

• The ASC achieves a cloud cover measurement accuracy of ±2% for cloud fractions below 20% and ±5% for cloud fractions above 80%, as validated with artificial cloud simulations.
• The system's night sky brightness measurements are calibrated against a Unihedron SQM-LE, enabling accurate conversion of pixel intensity to mag/arcsec² in the visible spectrum.
• The sensitivity threshold for star detection reaches magnitude 6 at zenith, with a 4-magnitude attenuation at 60° zenith angle due to combined optical and atmospheric effects.
• The algorithm fails when the lens is obscured by rain, snow, fog, or bird droppings, and cannot operate during moonlit conditions due to image saturation.
• The ASC system can operate autonomously for months without maintenance, with minimal operational costs, making it ideal for remote deployment at candidate observatory sites.
• The calibration process, validated by star position matching, ensures reliable transformation of pixel coordinates to azimuth and elevation across the full 185° field of view.