Uwarp

Or try an example

Method

Convolution mask size

Strength

Help Us Improve

—

(—)

Edge detection online

Detect edges in images online: Roberts, Sobel, or Laplacian operators, 3×3 or 5×5 mask, strength slider, live preview, PNG download in browser.

What is image edge detection?

Image edge detection finds boundaries where brightness or color changes sharply—outlines of objects, text strokes, and texture transitions. This page convolves a grayscale version of your photo with Roberts cross, Sobel-Feldman, or Laplacian kernels, then scales the gradient magnitude by strength to produce a bright-on-dark edge map. Mask size controls how far apart samples are: 3×3 for fine lines, 5×5 for smoother, wider edges. Upload, tune options, preview beside the original, and download PNG—all in your browser.

Edge detection features

Convolution-based border detection in the browser.

  • Three operators

    Roberts cross, Sobel-Feldman, and Laplacian edge filters.

  • Mask size choice

    3×3 for fine detail or 5×5 for broader edge lines.

  • Strength slider

    Scale gradient magnitude from subtle outlines to bold borders.

  • Automatic preview

    Output refreshes as you change method, mask, or strength.

  • Side-by-side panels

    Compare the original photo and edge map before download.

  • PNG download

    Save as image-edge-detection.png.

How to find edges in a photo

Steps for online image edge detection.

  1. Upload your image

    Drop or click in the input panel (15 MB max).

  2. Pick a detection method

    Try Sobel first; switch to Roberts for thin lines or Laplacian for fine change.

  3. Set mask and strength

    Use 3×3 for crisp detail; raise strength until edges read clearly.

  4. Download PNG

    Save the edge map for tracing, compositing, or further editing.

Tips for edge maps

Cleaner borders and less noise.
  • 01

    Start with Sobel

    Sobel-Feldman is a reliable default for most photos.

  • 02

    Reduce noise first

    Grainy sources amplify speckle—try /blur-image lightly if needed.

  • 03

    Use Roberts for thin lines

    Roberts cross works well on high-contrast logos and sketches.

  • 04

    Watch Laplacian noise

    Laplacian is sensitive; lower strength on smooth gradients.

  • 05

    Try 5×5 on small images

    Wider sampling can reduce jagged edges on low-res uploads.

  • 06

    Export PNG for overlays

    Multiply or screen the edge map in your editor for stylized looks.

Edge Detection Tools Compared — Which One Fits Your Workflow

How this free edge detection tool stacks up against other popular online tools for detecting image edges.

  1. PineTools.com

    Edge detection with Canny and Sobel filters. Basic upload-apply. No intensity control.

  2. ImageOnline.co

    Edge detection with adjustable threshold. Simple but no algorithm selection.

  3. Gamitisa.com

    Edge detect among 11 photo filters. 1 MB file size limit.

  4. Adobe Express

    Edge filters in full photo editor. Not a dedicated edge detection tool.

  5. Edge Detection Online — Find Image Edges

    Apply edge detection filters (Canny, Sobel, or Prewitt) with adjustable sensitivity. Real-time preview. Side-by-side comparison. All in browser.

When to use edge detection

Typical uses for image border finders.

  • Line art extraction

    Generate outlines from photos before manual tracing or vector work.

  • Vision prototyping

    Preview how gradient operators respond to shapes and textures.

  • Photo analysis demos

    Compare Roberts, Sobel, and Laplacian on the same source image.

  • Compositing masks

    Use high-contrast edge maps as starting points for selections.

  • Technical illustrations

    Highlight object boundaries for diagrams or documentation.

  • Pre-sharpen checks

    See where contrast concentrates before using /sharpen-image.

Why use this edge detector

Benefits of browser-based edge detection.

  • Operator choice

    Match the filter to your subject without installing OpenCV.

  • Private

    Images stay on your device—no server upload.

  • Live tuning

    Adjust strength and mask size with instant feedback.

  • Free

    Unlimited previews and downloads.

  • No install

    Works in any modern browser with Canvas 2D.

  • Alpha preserved

    Transparency on PNG uploads is kept in the output.

Technical details

How convolution edge detection works here.
Roberts cross
Diagonal 2×2 gradients; step 2 when 5×5 mask is selected.
Sobel-Feldman
Standard 3×3 Gx/Gy kernels; magnitude = hypot(gx, gy).
Laplacian
4-center minus cross neighbors; absolute response.
Strength scaling
Gradient magnitude × (strength / 100) × 2.5, clamped 0–255.
Input limits
15 MB; longest edge 8192 px.
Rendering
Canvas 2D getImageData / putImageData; grayscale RGB output.

Sources & References

Authoritative references on edge detection algorithms, convolution operators, and the browser technology this tool uses.
  • “Edge detection includes a variety of mathematical methods that aim to identify points in a digital image at which the image brightness changes sharply.”
    — Wikipedia — Edge detection

    Edge detection in computer vision: Edge detection identifies object boundaries by finding sharp intensity changes. This tool implements Roberts cross, Sobel-Feldman, and Laplacian operators — three classic edge detection methods. en.wikipedia.org/wiki/Edge_detection

  • “The Canvas API provides a means for drawing graphics via JavaScript and the HTML canvas element.”
    — MDN Web Docs — Canvas API

    Canvas API (client-side convolution): This tool performs convolution-based edge detection using Canvas getImageData and putImageData for pixel-level processing. MDN documents the pixel manipulation APIs used. developer.mozilla.org/en-US/docs/Web/API/Canvas_API

  • “PNG supports lossless data compression, making it useful for storing graphics that need to be edited repeatedly without losing quality.”
    — W3C — PNG Specification (overview)

    PNG image format: Edge map results download as PNG to preserve fine line detail without compression artifacts. PNG's lossless encoding is ideal for grayscale edge output. w3.org/TR/PNG/

  • “The Sobel operator, sometimes called the Sobel-Feldman operator or Sobel filter, is used in image processing and computer vision, particularly within edge detection algorithms.”
    — Wikipedia — Sobel operator

    Sobel operator for gradient-based edge detection: The Sobel-Feldman operator uses 3×3 convolution kernels to approximate gradient magnitude. It is one of the three methods available in this tool and often the recommended default. en.wikipedia.org/wiki/Sobel_operator

Frequently Asked Questions

Quick answers about this tool—open a question to read more.