How to Read Morse Code from a Picture

To read Morse code from a picture, isolate a straight row of dots and dashes, preserve the gaps, and compare the detected code with the image before using the translation. The image reader can make a first pass, but a border, caption or missing letter gap can change the result. The examples below use downloadable files so you can repeat each step.

Start with the original image

Choose a PNG or JPG, or paste a screenshot into the image panel. A screenshot is often clearer than a camera photo of the same screen. Keep an untouched copy: if a crop removes the start of a word, restoring the original is easier than trying to reconstruct it. The reader processes your file on your device and does not upload it.

Look for short and long marks arranged in a row. A dot is short; a dash is normally about three dot lengths. The spaces matter just as much. If every mark is equally spaced, the picture may show an unbroken signal rather than individual letters. Decorative dots and lines are not necessarily Morse.

Crop a screenshot with a caption or border

Open the bordered test image. Its intended message is MEET AT NOON, but the default reader mistakes the frame and printed caption for signal. This is a synthetic control, not a customer photograph. The test manifest records the wrong default output and the separate result after cropping.

1. Drag a rectangle around the horizontal strip in the left preview. Leave some white space around the marks.
2. Choose Apply crop. On a small touchscreen, open the crop-edge controls instead: use the percentages printed beside the bordered case in the test set.
3. Check that the first dot and final dash are still inside the crop. A tight crop can remove part of a letter.
4. Compare the new detection boxes with the strip. This control reads MEET AT NOON after the caption and frame are removed.

If your real image has several rows, first test one row. Separate rows may represent different words, letters or even columns. The reader treats a new row as a word break, which may not match the author's layout. Restore the image and crop another region if you need to read it separately.

Rotate or straighten before changing the threshold

The 90-degree test image is a vertical version of the same strip. In its default orientation the reader produces repeated E characters. Rotate left 90 degrees and it reads MEET AT NOON. The rotation changes the layout the reader sees; increasing the threshold cannot turn a vertical column into a horizontal row.

For a slight tilt, open the straighten controls, choose an angle and apply it. Try a small angle in the opposite direction to the tilt. Keep all the marks visible and inspect the result. Perspective distortion, where one end of a row is much smaller than the other, needs a better photograph or an editor that can correct perspective. Rotation alone will not repair that.

Check polarity and contrast

The initial read uses an automatic threshold and chooses light or dark marks. The inverted control and low-contrast control both read MEET AT NOON with automatic settings. If light marks disappear from the detection preview, try Invert. Move the threshold only after cropping and rotation are correct.

Watch the preview while moving the slider. A useful setting keeps each mark solid without joining neighbouring marks to shadows or background texture. If the marks vanish at one setting and merge at the next, take a clearer picture or transcribe the code manually. A high detection check measures consistent mark shapes; it is not a promise that the words are right.

Locate a wrong letter in the picture

Green boxes belong to recognised letter groups. Red boxes belong to groups that have no mapping in the ordinary-letter table. Tap a box to select its group in the detected Morse field. Check whether a dot was missed, a dash was broken, or two letters were joined. Then edit that group. The English result updates as you type.

Manual corrections do not change the original pixel detection. Once you edit the Morse field, the boxes still show the original read. Adjusting an image control runs detection again and replaces the edited transcription, so copy a useful correction before doing that. Save result card exports the current transcription and translation locally, without your original image.

Recover letter and word boundaries

In typed Morse, one space separates letters; / or three spaces separates words. .- is A, while . - is E followed by T. An image needs visible equivalents of those boundaries. In ordinary Morse timing the gaps are one unit inside a letter, three between letters and seven between words.

When letter gaps are missing, several messages may fit the same stream. The tool lists possible splits for short streams; those are alternatives, not a guessed solution. A Morse discussion about formatting losing gaps shows the practical problem, and another discussion about merged dashes describes how a font can change the marks. These are examples of reported difficulties, not endorsements of this reader.

The independent SOS diagram in the test set deliberately has no letter gaps: it represents the known continuous SOS prosign ...---.... An ordinary-letter decoder shows an unrecognised group until you interpret that context or separate S, O and S. Do not assume that every unfamiliar unbroken pattern is SOS.

When manual reading is the better choice

The independent Metadata diagram uses a stylised layout, and the outdoor V photograph contains large sculptural forms against a busy background. Their default outputs in the test set are wrong. For an unconventional design, compare each intended mark with the alphabet chart, type your transcription into Morse to English, and preserve uncertainty about missing boundaries.

If you are writing a clue, use the puzzle kit to create and verify an exported PNG or audio file. For classroom practice, the worksheet and answer key give learners examples with explicit gaps. A useful final check is simple: can a second reader reach the same message from the final file without being told the answer?