DetectMethod

Specifies the method by which corners are detected.

Remarks

The 'Detect Method' enumeration determines which algorithm is used by Corner Detection to identify corner points.

• Use Square for structured forms, tables, or documents where corners are expected to be at right angles (90°).
• Use Harris for general-purpose corner detection, including rotated, skewed, or natural layouts where corners may be at arbitrary angles.

Practical Guidance

• Choose the method that best matches the geometry of your documents.
• Review diagnostic overlays to compare the results of each method.
• Tuning parameters such as threshold, window size, and mask size may have different effects depending on the selected method.

Can be one of the following values:

Name	Value		Description
Square	0	►	Detects non-skewed 90° corners in an image. The Square method is optimized for detecting corners where two lines meet at a right angle, such as the corners of boxes, tables, or form fields. Best for structured documents with grid-like layouts. Less effective for rotated, skewed, or natural images. Tuning 'Mask Size', 'Border Size', and 'Min Confidence' can help target the expected scale and location of corners.
Harris	1	►	Detects arbitrary corners at any angle. The Harris method uses the Harris corner detection algorithm to find corners at any angle, making it suitable for a wide range of document types. Best for general layouts, rotated or skewed documents, and natural images. Sensitive to noise; tuning 'Corner Threshold', 'Window Size', 'Filter Type', and 'K Parameter' is important for optimal results.

Used By

Corner Detection