Sunset Radiance FieldEarth radius instrument
Earth radius— km
Local image analysis
Calibrated horizon-dip method

Measure Earth from one photograph.

Upload a clear sea horizon. The tool finds the line and reads useful photo metadata; you answer the remaining questions in plain language.

Calculated resultR = — km
Attach a clear sea-horizon photo

No known Earth-radius value is used as an input.

01 / Image analysis

Visible horizon

02 / Three answers

Complete the measurement

0/3 ready
Was the camera level when the photo was taken?
“Assume level” is valid only when level was verified while taking the photo. A small unnoticed tilt can turn the answer into kilometres instead of thousands of kilometres. Positive tilt points upward; negative points downward.
  • Horizon extracted from the photo
  • Camera height supplied
  • Lens calibrated
  • Camera level answered
Advanced precision and refraction

Optional. Add known measurement precision to produce a more honest radius range.

The result is computed from your height and the angular dip between true horizontal and the detected horizon. It is never replaced with the accepted Earth radius. A broad Earth-scale sanity check rejects broken calibration instead of presenting it as Earth's radius.

Measurement boundary

The photo supplies the visible horizon and radiance field. Height, lens calibration, camera level, lens distortion, and atmospheric refraction cannot always be inferred from pixels. A level assumption is clearly marked because it directly affects the result, and a non-Earth-scale output is rejected as failed calibration.

How to use Sunset Radiance Field

  1. Choose a suitable photograph

    Attach an uncropped image with a clear, unobstructed sea horizon. The tool analyses a local copy scaled to at most 1,200 pixels wide.

  2. Review the detected horizon

    Check the amber line against the visible horizon. Drag it vertically or use the position control if automatic detection selected the wrong edge.

  3. Calibrate the camera geometry

    Enter camera height above sea level and the 35 mm-equivalent focal length, then confirm that the camera was level or provide its known tilt.

  4. Add uncertainty when known

    Optionally enter measurement precision and a refraction coefficient so the reported distribution reflects those assumptions.

  5. Calculate and interpret

    Run the calculation, inspect the horizon dip and uncertainty range, and use the sensitivity views to see which inputs most affect the result.

Formats, compatibility, and limits

Inputs

  • A local image file accepted by the browser, ideally an uncropped photograph of a clear sea horizon.
  • Camera height above the sea in metres.
  • A 35 mm-equivalent focal length in millimetres.
  • A verified level-camera assumption or a known camera tilt in degrees.
  • Optional uncertainties for height, field of view, tilt, and horizon position, plus an optional refraction coefficient from 0 to 0.5.

Outputs

  • An estimated Earth radius in kilometres and the measured horizon-dip angle.
  • A sampled radius distribution with median, 68% range, and 95% range when valid inputs are available.
  • On-page horizon, radiance, atmosphere, uncertainty, and sensitivity visualisations; the current tool does not export a report file.

Compatibility

  • Runs in modern desktop and mobile browsers that can decode the selected image and read Canvas 2D pixels.
  • A larger screen makes manual horizon review and the scientific visualisations easier to inspect.
  • EXIF autofill is optional; unsupported or missing metadata can be entered manually.

Known limits

  • Accuracy is highly sensitive to camera height, focal-length calibration, unnoticed camera tilt, horizon placement, lens distortion, and atmospheric refraction.
  • The pinhole-camera geometry assumes the supplied 35 mm-equivalent focal length describes the original uncropped image.
  • Automatic horizon detection can select a foreground edge, cloud boundary, or shoreline obstruction and must be reviewed by the user.
  • Radiance contours and scattering curves are explanatory views only; brightness is not inverted to obtain the radius.
  • A broad sanity range from 1,000 to 20,000 km rejects obviously broken calibration, but passing that check does not validate the estimate.

Privacy and processing

  • The selected photograph, its pixels, and supported EXIF fields are analysed in the browser; this tool does not upload them to an application server.
  • The image is held in page memory through a temporary object URL and is not written to local storage or IndexedDB by the tool.
  • Replacing the image or leaving the page releases the temporary image reference; refreshing also clears the measurement state.

How the calibrated horizon-dip calculation works

The calculation derives a vertical field of view from the photograph’s aspect ratio and 35 mm-equivalent focal length. It maps image rows through a pinhole-camera projection and measures the angle between true horizontal and the reviewed horizon. Camera height and that angular dip produce the spherical-radius estimate, with an optional correction for refraction. For uncertainty, a deterministic 4,096-sample calculation varies the measurements you marked as uncertain. The accepted Earth radius is never substituted into the result. A deliberately broad Earth-scale check only blocks implausible calibration.

Questions about Sunset Radiance Field

Does the calculator use the accepted Earth radius to produce its answer?

No. The numeric result comes from camera height and the measured angular horizon dip. The tool only applies a broad Earth-scale range to reject clearly broken calibration instead of presenting it as a valid Earth measurement.

Does the sunset brightness or radiance field determine the radius?

No. Bright-pixel contours and atmospheric curves are explanatory visualisations. The radius calculation uses the reviewed horizon position, camera height, lens-derived field of view, camera tilt, and optional refraction.

How accurate is an Earth-radius estimate from one photograph?

The estimate can be far off if the camera was not demonstrably level, or if height, focal length, distortion, refraction, or horizon placement are uncertain. Treat it as an educational experiment and read the uncertainty and sensitivity panels alongside the result.

What kind of photograph works best?

Use an original, uncropped image with a long, sharp, unobstructed sea horizon. Avoid nearby land, strong foreground edges, heavy distortion, and images whose focal length or camera orientation cannot be recovered.