> 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/detectors/shape-detector.md).
# Shape Detector
This function block finds a distinct shape in a color image by filtering, thresholding and contour analysis. It returns a visualized result and numeric shape descriptors that can be used by downstream blocks.
## đĨ Inputs
`Image` This input socket accepts a color image to analyze.
## đ¤ Outputs
`Filtered Image` Output socket that provides the binary or filtered image used for contour extraction (visual feedback for preprocessing).
`Contoured Image` Output socket that provides the original image with detected contour drawn.
`Shape Image` Output socket that provides the cropped or separated shape visualization.
`Border Coordinate` Output socket that provides the contour coordinate(s) that form the detected border.
`Shape Width` Output socket that provides the detected shape width.
`Shape Height` Output socket that provides the detected shape height.
`Shape Position` Output socket that provides the detected shape center position.
`Shape Rotation` Output socket that provides the detected shape rotation angle.
## đšī¸ Controls
`Blur coefficient` Slider control to tune median blur strength before thresholding. Use to reduce noise while preserving edges.
`Clear spots parameter` Slider control to set morphological closing size. Use to remove small holes or join object parts.
`Auto threshold ?` Checkbox control to enable automatic thresholding (Otsu). When enabled, manual threshold range is disabled.
`Threshold range` Two-ended slider control to set manual threshold low/high values. Active when `Auto threshold ?` is unchecked.
`Shape area range %` Two-ended slider control to limit detected contours by relative area (percentage of image). Use to ignore very small or very large objects.
## âī¸ Running mechanism
* When an image is provided to the `Image` input socket, the block applies the selected blur, thresholding (auto or manual) and morphological closing.
* Contours are extracted from the filtered image. The block selects contours that fall into the configured `Shape area range %` and returns the largest valid contour and related outputs.
* Visual outputs show the filtered image and the detected contour drawn on the original image so you can verify results in the UI.
* Toggling `Auto threshold ?` instantly enables or disables the manual `Threshold range` control to let you experiment quickly.
## ⨠Features
* Visual debugging outputs for both the thresholded image and contour overlay.
* Flexible preprocessing: median blur and morphological closing parameters.
* Area-based filtering to avoid false contours from noise or borders.
* Produces both visual outputs and numeric descriptors (size, position, rotation) for downstream logic or logging.
## đ Usage instructions
1. Feed a color image into the `Image` input socket.
2. Start with moderate `Blur coefficient` and a small `Clear spots parameter`.
3. Try `Auto threshold ?` first. If results are noisy, uncheck and tune `Threshold range` manually.
4. Adjust `Shape area range %` so the desired object falls within the allowed area window.
5. Inspect `Filtered Image` and `Contoured Image` to confirm detection. Use downstream blocks to act on `Shape Position` or `Shape Rotation`.
## đĄ Tips and Tricks
* If the image has color noise or lighting variations, try adding a preprocessing chain before this block with `HSV Filter` or `Adjust Colors` to improve separation of the object from background.
* Use `Blur` or `Denoising` upstream to reduce speckle before thresholding for more stable contours.
* If you need to analyze only a sub-region, crop first with `Image ROI Select` or `Image ROI` and feed the cropped image into this block.
* To convert a noisy color scene into a clearer binary input, combine with `Image Adaptive Threshold` or `Image Threshold` before this block.
* For further geometric processing or feature extraction, connect outputs to related shape analysis blocks such as `Approximate Contour`, `Minimum Rectangle`, `Minimum Rotated Rectangle` or `Translate Shape`.
* Visualize results live using the `Show Image` block and save examples or logs with `Image Logger` or `Image Write` for later inspection.
## đ ī¸ Troubleshooting
* No contour found: increase the upper bound of `Shape area range %` or lower the minimum; check `Filtered Image` to confirm the object appears after thresholding.
* Too many small contours: increase `Clear spots parameter` (morphological closing) or tighten the minimum area in `Shape area range %`.
* Object split into multiple contours: reduce blur or decrease morphological closing, or crop region with `Image ROI Select` to isolate the object.
* Threshold not working: try enabling `Auto threshold ?` to let the block pick a suitable threshold automatically, then fine-tune the manual `Threshold range` if needed.