CES 2027 · LAS VEGAS 2027.01.06–09 · LVCC South Hall 2 · Booth 35915 D-65 Exhibition info · Book a meeting →
KOENDE中文

Insights · AI Visibility

How AI visibility meters work: from extinction coefficient to MOR

· SIJUNG Inc.

An AI visibility meter measures visibility by analyzing images from a dedicated camera with AI. This article explains the extinction coefficient, meteorological optical range (MOR) and prevailing visibility, which underpin low-visibility decisions in fog and sea fog, together with how video-based visibility measurement works and how it is applied on site.

Visibility and meteorological optical range (MOR)

Visibility is the greatest distance at which a black object of suitable size can be recognized by day, or a light of a given intensity by night. It is defined by the World Meteorological Organization (WMO) and serves as the reference for aviation, road and maritime safety decisions.

What an instrument actually reports is the meteorological optical range (MOR): the path length over which a parallel beam from an incandescent lamp is attenuated to a defined fraction as it passes through the atmosphere. MOR is inversely proportional to the atmospheric extinction coefficient (σ).

The extinction coefficient expresses how much light is attenuated per unit distance by scattering and absorption from aerosols and fog droplets. As fog or sea fog thickens, particle scattering increases, the extinction coefficient rises and conditions turn to low visibility.

How conventional visibility meters measure: point and line measurement

A forward-scatter visibility sensor measures scattered light in a small sampling volume in front of the sensor, which is a point measurement. A transmissometer measures the attenuation of light between a transmitter and a receiver, which is a line measurement.

Both are proven methods used in international observation, but each represents only the air at one spot or along one path. Sea fog that forms and clears locally within a short time, or fog whose density differs by direction, is hard to capture from numbers alone. With no images stored alongside the readings, after-the-fact verification is also limited.

How video-based visibility works: estimating the extinction coefficient by image analysis

An AI visibility meter captures its whole field of view, referenced to targets whose distances are surveyed at the installation site, and estimates the atmospheric extinction coefficient along the observation path by image analysis. What is observed is the actual atmospheric path the camera looks along, not the air at a single spot.

The estimated extinction coefficient is converted to MOR to give the visibility value. Visual estimation of distance by an observer is replaced with images and AI judgment, so observations can be repeated at any time on the same basis.

Because many targets in the field of view are analyzed at once, visibility by distance and direction is obtained in a single pass, which is an area measurement. The omnidirectional model JS-08 derives prevailing visibility from the visibility in each direction.

On-device processing and visibility nowcasting

Everything from image capture to extinction coefficient estimation and MOR calculation is processed on the device. Readings are stored together with evidence images, so the reason behind each visibility value can be checked later in the field images.

Visibility nowcasting is a core function that forecasts upcoming visibility from the trend in the visibility time series. Early warnings issued before low-visibility thresholds are reached support operational decisions in airport, bridge and port control.

Field application and intercomparison at a standard observatory

To ensure measurement accuracy, the system is supplied as a dedicated camera together with analysis equipment rather than connected to existing CCTV. Target distances are surveyed and the system is calibrated at the installation site.

At the KMA Gochang Standard Weather Observatory, ST-01 is installed in the same observation row as leading visibility meters from around the world, including a reference transmissometer, to compare readings under the same atmosphere at the same time (intercomparison). The test builds data to confirm the observational reliability of video-based visibility measurement against international standards.

  • Airports and forest aviation: omnidirectional visibility around runways and flight routes (JS-08)
  • Bridges and roads: low-visibility monitoring and early warning on fog-prone sections (ST-01)
  • Ports and maritime traffic control: monitoring sea fog formation and visibility changes
  • Weather observation: video-based visibility combined with weather elements (JS-02P)

AI visibility product line →Product details →Case studies →All insights →

Contact

Request a site review

After reviewing the installation site, observation range and detection targets, our engineers propose a configuration suited to the site

Contact us +82-42-933-5520