Edge Detection
Find the edges in a picture with a Sobel or Canny filter (white outlines on black, black on white, or drawn over the photo) and save the edge map at full resolution.
or browse filesdrop an image here or paste
.png
Find the edges in a picture with a Sobel or Canny filter (white outlines on black, black on white, or drawn over the photo) and save the edge map at full resolution.
or browse filesdrop an image here or paste
This free edge detector finds the outlines in a picture and gives you the edge map as an image. Open a photo and the Sobel filter runs straight away: it measures how quickly brightness changes at every pixel and draws the result as grey edges, bright where the contrast is strong. Switch to Canny for the classic computer vision detector - Gaussian smoothing, gradient, non-maximum suppression and hysteresis - which thins every edge to a single pixel and keeps a faint edge only where it connects to a strong one. One Sensitivity slider drives both: low picks out only the boldest contours, high pulls in fine texture.
Sobel is the fast, soft detector: it responds in proportion to contrast, so the edge map keeps a sense of which boundaries matter and works well as an artistic effect, a relief look or a base for a sketch. Canny is the precise one: thin, connected lines with no double edges, the output you want for tracing, for measuring, for a colouring-page outline or for feeding another algorithm. The details card on the picture reports the method, the thresholds in use and the share of pixels marked as edge, which is the number to compare when tuning a detector for a computer vision project.
The default map is white edges on black, the way OpenCV and most textbooks show it. Black on white gives a line drawing that prints on paper and works as a stencil or a colouring page. Over the photo draws the detected edges in red on top of the original, which shows at a glance what the detector found and missed. Whichever you choose, the download is rendered from the full-resolution original, so a 4000 × 3000 photo gives a 4000 × 3000 edge map with every one-pixel Canny line intact. For a pencil look rather than a technical edge map, a sketch effect shades the picture instead of outlining it; this page stays a detector.
Preparing an image for tracing or laser engraving, checking focus by looking at how crisp the edges are, extracting the outline of a logo or a product, building a coloring page from a photograph, generating contours for a design, or simply understanding what a Sobel or Canny filter does before writing one. Because the whole detector runs in your browser, it is also a quick way to try threshold values on your own images before setting them in code.
Nothing is uploaded. The photo is decoded, filtered and encoded on your device, with no signup, no watermark and no file size limit beyond your own memory.
Open the picture; the Sobel edge map appears immediately. Adjust Sensitivity, choose Canny for thin one-pixel edges, pick the output style, and press Download to save the map at full resolution.
Sobel computes the brightness gradient at each pixel and shows its strength, so edges are soft and proportional to contrast. Canny smooths the image, thins each edge to one pixel at the gradient's peak and keeps weak edges only when they connect to strong ones, giving clean connected outlines.
For Sobel it scales the gradient, so higher values make faint edges visible. For Canny it lowers the high threshold, with the low threshold following at 40% of it, so more of the weaker edges are kept. The thresholds in use are shown in the details card.
Yes. Choose Black on white in the Output row; the same edge map is inverted, which is the form for printing, stencils and colouring pages.
Yes. The detector runs again on the original file when you download, so the saved map has the photo's full pixel size rather than the preview's.
No. Detection runs entirely in your browser and the file never leaves your device.