Detect Edges
Function
This Step extracts object edges from the input 2D image to highlight their contour features for subsequent analysis or processing.
Usage Scenario
This Step is suitable for scenarios where object edges need to be extracted from images, such as object contour extraction, geometric fitting, edge positioning, etc. It is a general 2D image edge detection Step and has no fixed usage scenarios.
Parameter Description
| Parameter | Description |
|---|---|
Segment Type |
Description: This parameter is used to select a specific edge detector. The edge quality, noise immunity, and accuracy of different detectors vary, and you should select them according to actual requirements. Value list: CannyDetector, EdgeDrawingDetector, SobelDetector Default value: CannyDetector |
CannyDetector |
Description: This type of detector has strong noise suppression ability, high edge positioning accuracy, and can generate clear and continuous edges. It is suitable for scenarios with noise but clear edges.
Instruction: Edges with gradients above the high threshold are considered actual edges. Edges with gradients between high and low thresholds are considered actual edges only if they are connected to actual edges with gradients greater than the high threshold. Edges with gradients below the low threshold are ignored. |
EdgeDrawingDetector |
Description: This type of detector detects edges with strong continuity, which is suitable for scenarios that require high edge integrity and where coherent profiles need to be obtained. |
SobelDetector |
Description: This type of detector is fast at calculating and suitable for coarse edge locating in high-quality images and regular contours. The output edge is wide, providing continuous and clear visual effects when the image quality is good.
Instruction: A smaller kernel size preserves more detail; a larger kernel size enhances smoothing and helps reduce noise, but may lead to a weakening of edge details. |