ImgHeal
ImgHeal photo guide

How to Check Bicycle Spokes in a Restored Photo

By ImgHeal · Updated

Review wheels, spokes, and motion cues so a restored bicycle still has believable geometry.

Establish the wheel perspective

Inspect the outside of each wheel before considering individual spokes. A wheel viewed at an angle appears elliptical, and the two wheels may present different shapes because of steering. Do not force both into perfect circles. Note the axle locations and where the wheels meet the ground; those relationships help reveal edits that subtly warp the bicycle.

Allow for movement and soft detail

Moving spokes may form a blur rather than separate lines. A bright arc or faint fan can be a real motion cue, not damage to remove. Check whether the frame and background are sharper than the wheels. That difference helps explain the source without demanding a full set of spokes the camera never resolved.

Check hub-to-rim connections

In the restored version, look for spokes that attach outside the hub, cross the tire, or become repeated decorative patterns. Inspect the fork and mudguards at the same time. If a scratch needs manual attention in a separate editor, use the surviving geometry on either side and leave uncertain fine details soft rather than redrawing an entire wheel.

Keep the ride believable

View the full photograph at its intended size. A very sharp wheel can look detached from a motion-blurred rider or road. Compare the tire contact points and bicycle lean with the source. Save a restrained result if added precision changes the feeling of movement, and retain the original scan when the bicycle itself is an important family object.

Check this scene before saving

  • Check perspective before circularity.
  • Preserve genuine wheel motion blur.
  • Inspect spoke connections and ground contact.

Continue with a related guide

Try a restoration on a copy of your photo

Keep the original scan. Upload a separate digital copy, review ImgHeal’s generated result, and decide whether it suits your needs.

Open ImgHeal