
Treat ImageMagick Canny's lower and upper thresholds as different edge-detection controls. Hold the Gaussian smoothing setting fixed while comparing them, then inspect the actual edge map. Raising a threshold is not a generic way to sharpen a photograph or a guarantee that every remaining line is meaningful.
Define the required edge artifact
Choose the particular boundaries that must remain visible in the output, such as the outline of a simple diagram. Identify a short meaningful line and an unwanted weak texture separately.
The official Canny reference documents percentage thresholds with the lower value below the upper value. Its example 0x1+10%+30% includes Gaussian smoothing before the two thresholds.
Establish one baseline
A proposed POSIX-shell starting point is:
magick input.png -canny '0x1+10%+30%' edges-baseline.png
This command has not been executed. Its values are a documented example, not image-specific recommendations. Confirm the installed ImageMagick 7 build and adapt percent-sign quoting to the actual shell, especially if translating to a Windows batch file.
Record the source and all settings. Changing the smoothing sigma along with the thresholds makes it harder to determine which choice shortened or removed an edge.
Compare component selection and extent
The manual describes increasing the upper threshold, with the lower fixed, as reducing detected edge components while retaining their lengths. Raising the lower threshold, with the upper fixed, shortens their extent. Use those documented roles to organize controlled comparisons rather than moving both numbers together.
Inspect the short required line, the longer outline and unwanted texture in each actual result. Do not infer that a lower component count is automatically cleaner if a required boundary disappears.
Keep analysis metrics and line geometry separate
The FFmpeg blurdetect guide uses Canny responses inside a blur measurement. Its final artifact is a reading, not the standalone edge raster produced here.
Likewise, the cropdetect guide estimates picture bounds, rather than preserving specified diagram edges. Canny output alone neither calculates physical line dimensions nor supplies a complete object segmentation.
Accept the edge-map candidate when the required boundaries remain useful and unwanted responses are tolerable for its next operation. This is a proposed comparison procedure with no measured detection accuracy, performed conversion or guaranteed recovery of weak source detail.
Related tasks
- How Do FFmpeg blurdetect’s Edge Thresholds and Block Options Change the Analysis?
- How Do FFmpeg cropdetect’s Black and Mvedges Modes Decide the Picture Area?
Sources and scope
- ImageMagick 7 option reference: canny: Canny uses ordered percentage thresholds and separate Gaussian smoothing; the two thresholds affect detected components and their extent in different ways.
Official ImageMagick 7 documentation was captured October 23, 2026. Examples, calculations and inspections are proposed rather than executed. Confirm installed-build support and shell interpretation before use.
Prepared by Imagild Editorial, the company editorial team responsible for these guides. AI assisted research and drafting. This is documentation-based guidance with proposed examples, not a claim of hands-on testing or an independent product review. Report a correction.
Cover: Pignatta / Pixabay, Pixabay Content License. This is an illustrative source photograph, not an application screenshot or a generated test result.