> For the complete documentation index, see [llms.txt](https://docs.augelab.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.augelab.com/function-blocks/blocks-reference/detections-shapes/shape-analysis/approximate-contour.md). # Approximate Contour This function block simplifies a given contour by creating an enclosing polygon with a reduced number of edges. It visualizes both the original contour and the simplified approximation, and provides geometric information such as center, perimeter and corner count. ## đŸ“Ĩ Inputs `Image Any` The image used for visualization (draws original and approximated contours on this image). `Contour` The input contour to be approximated (shape data). ## 📤 Outputs `Image Any` Image that shows the original contour and the approximated polygon with corner markers. `Contour` The approximated contour as a shape output. `Center` The centroid of the approximated contour (x, y). `Perimeter` Perimeter length of the approximated contour. `Number of corners` Number of vertices in the approximated polygon. ## 🕹ī¸ Controls `Epsilon` Slider to set the approximation tolerance. Higher values produce simpler polygons with fewer corners; lower values keep more detail. ## ⚙ī¸ Running mechanism * When the block runs, it receives an image and a contour through its input sockets. * The block uses the `Epsilon` control to determine how aggressively the contour is simplified (epsilon is applied relative to contour perimeter). * The block creates a simplified polygon and draws both the original contour and the approximation onto the provided image for visual feedback. * The block then calculates the approximated contour's centroid, perimeter and corner count and sends these values through the output sockets. ## đŸŽ¯ Features * Visual comparison: original contour (one color) and approximated polygon (another color) are both shown on the output image, with corner points marked for quick inspection. * Numeric outputs: center position, perimeter and number of corners are provided for downstream logic or logging. * Adjustable approximation: the slider gives an intuitive way to balance between simplification and fidelity. ## 📝 Usage instructions 1. Provide a contour (for example from a contour-finding block) to the `Contour` input. 2. Provide the image you want to annotate to the `Image Any` input. 3. Adjust `Epsilon` until the approximated shape matches the desired level of detail. 4. Use the outputs for visualization, measurement or further processing. ## 💡 Tips and Tricks * To obtain contours suitable for approximation, use the `Find Contour` block first. That block detects contours you can feed into this block. * Preprocess the source image with `Blur`, `Image Threshold` or `Morphological Transformations` to reduce noise and produce cleaner contours. * Use `Image ROI` or `Image ROI Select` to crop the region of interest before contour detection and approximation to speed up processing and avoid irrelevant shapes. * After approximation, you may want to: * Convert the approximated contour into an image crop with `Contour to Image` for focused inspection. * Fit simple enclosing shapes using `Minimum Rectangle`, `Minimum Circle` or `Minimum Ellipse` for measurements or alignment tasks. * Fill the approximated shape with `Fill Contour` when you need a binary mask of the region. * Use `Draw Rectangle`, `Draw Point` or other draw blocks to overlay additional annotations or combine with detection results. * To shift shapes in coordinate space (for ROI adjustments or alignment), use `Translate Shape`. ## 🛠ī¸ Troubleshooting * If the approximated polygon is too coarse or too detailed, tweak the `Epsilon` slider gradually until you get the desired corner count. * If contour outputs are empty or invalid, ensure the contour source (for example `Find Contour`) is producing valid shapes and that any preprocessing steps produce a clear foreground/background separation. * If visual markers are misplaced after cropping or coordinate transforms, verify that subsequent blocks (ROI, translation or drawing) use consistent coordinate systems.