2D image edges
Feature Description
This operator uses the classic Canny edge detection algorithm to extract edge features from the input image. It employs multi-stage processing to find locations of intense grayscale changes in the image, thereby effectively detecting edges while suppressing noise.
Use Cases
-
Feature Extraction: Extracting object contours as key features before tasks like shape recognition, object detection, or image matching.
-
Image Segmentation Aid: Edge information can serve as input or reference for image segmentation algorithms.
-
Visual Inspection: Detecting defects such as scratches and cracks on product surfaces, which usually manifest as distinct edges.
Inputs and Outputs
Input Item |
Image: The image for edge detection, can be grayscale or color. |
Output Item |
Result image: The edge image obtained after Canny algorithm processing. This is a binarized grayscale image where detected edge pixels are typically white, and background pixels are black. |
Parameter Descriptions
Low Threshold
Parameter Description |
Low threshold; if the pixel intensity change is below this value, it is definitely not an edge. |
Parameter Tuning Guide |
Lowering the threshold: May allow less obvious, weaker edges to be retained, making edge lines more complete and coherent, but may also misclassify some "noise" that is not originally an edge as an edge. Raising the threshold: Screens edges more strictly, only parts with strong changes will be retained, and weak edges will be ignored. |
Parameter Range |
[0, 1000], Default value: 1 |
High Threshold
Parameter Description |
High threshold; if the pixel intensity change is above this value, it will definitely be judged as an edge point. |
Parameter Tuning Guide |
Increasing the threshold: Will cause only very obvious edges to be detected, and the number of edges in the result may decrease; Decreasing the threshold: Will detect more weaker edges, the number of edges may increase, but may also include more false edges caused by noise or texture. |
Parameter Range |
[0, 1000], Default value: 100 |
Edge Detection Window Size
Parameter Description |
The aperture size of the Sobel operator used when calculating the image gradient, affecting the smoothness of the gradient calculation and sensitivity to noise. |
Parameter Tuning Guide |
|
Parameter Range |
[3,5,7] , Default value: 3 |
Precision mode
Parameter Description |
Switches between fast mode (L1 norm) or precise mode (L2 norm) for calculating gradient magnitude. Fast mode is used by default. |
Parameter Tuning Guide |
|