Contour

Filter

Sobel

Applies the Sobel operator – calculates gradient magnitude and highlights edges and intensity changes.

The Sobel filter highlights edges by measuring brightness changes in the image. It calculates gradients in the x- and y-directions and combines them to create an edge strength. This makes object outlines visible, robust against slight noise.

Flow

Parameter set

Parameter	Type	Description	Value range
Image	Image	Requires a grayscale image. This can be created by using the "Convert To Grayscale" node.
Order X	Int32	Specifies the gradient order in the x-direction.	0…2
Order Y	Int32	Specifies the gradient order in the y-direction.	0…2
Kernel Size (Pixel)	Int32	A high value increases the kernel size for calculating larger surface gradients. The size of the extended Sobel kernel must be 1, 3, 5, or 7.	1,3,5,7
Scale	Single	Scaling factor applied to the calculated gradient.	>0.1
Delta	Single	Value added to the calculated gradient.

Canny

Applies the Canny edge detection algorithm – detects edges based on intensity gradients in the image.

The Canny detector detects edges with exceptional precision through multi-stage processing (smoothing, gradient, non-maximum suppression, and hysteresis). It delivers thin, clear contours and reliably suppresses noise. This makes it suitable for precise edge analysis and complex scenes.

Flow

Parameter set

Parameter	Type	Description	Value range	Effect Min	Effect Max
Image	Image	Requires a grayscale image. This can be created by using the "Convert To Grayscale" node.
Threshold	Single	The value must be between 0 and 255. Higher values increase edge sensitivity, potentially detecting more edges.	0…255

Laplacian

Applies the Laplace operator – calculates second derivative and enhances areas with rapid intensity changes.

The Laplacian filter uses the second derivative and marks edges with sharp intensity changes. It detects edges regardless of their direction. Because it is sensitive to noise, it is often combined with prior smoothing.

Flow

Parameter set

Parameter	Type	Description	Value range
Image	Image	Requires a grayscale image. This can be created by using the "Convert To Grayscale" node.
Kernel Size (Pixel)	Int32	The value must be positive and odd. A high value increases the kernel size for calculating larger surface gradients.	1,3,5,7
Scale	Single	Scaling factor applied to the calculated gradient.	>0.1
Delta	Single	Value added to the calculated gradient.

Find Contours Basic

This node detects and outlines contours (white areas) in a binary image. Suitable inputs include the Binary Threshold, Adaptive Threshold, Color Threshold, Canny, or other methods that generate a binary image from a grayscale image. Three modes are available: Inner Contour, Outer Contour, or Both Contours – these allow the inner and outer contours of a surface to be detected. The detected contours can then be used for further processing.

Flow

Parameter set

Image preparation	Type	Description	Value range	Before	Thereafter
Image	Image → Region	Converting an image into a region using the Color Threshold node. The desired color is previously defined in the Make Color node—for example, yellow.
Image	Image → Region	The same conversion of an image into a region using the Color Threshold node. The desired color is previously defined in the Make Color node—for example, white.
Parameter	Type	Description	Value range	Before	Thereafter
Detection Mode	InnerContour	Detects only the inner contours of objects. This mode is useful when you want to focus on closed shapes within a binary region and ignore the outer boundary.
	OuterContour	Detects only the outermost contours of objects. This mode is suitable for isolating the outer boundaries of objects without considering nested structures.
	BothContours	Detects both the outermost contours and all nested inner contours. This mode is ideal for analyzing hierarchical relationships between contours, for example, when objects are enclosed by other objects.

Find Contour Advanced

This node detects and outlines contours (white areas) in a binary image. For it to work, the objects must be white – if necessary, the Invert Pixel node can be used for preprocessing. The detected contours are output sorted from largest to smallest and can be used for further processing.

Compared to the Find Contour Basic node, the Find Contour Advanced node allows you to control the process and thus the processing speed, as well as to specifically influence the sorting of the output contours.

Flow

Parameter set

Image preparation	Type	Description	Before	Thereafter
Image	Image → Region	Converting an image into a region using the Color Threshold node. The desired color is previously defined in the Make Color node—for example, yellow.
Image	Image → Region	The same conversion of an image into a region using the Color Threshold node. The desired color is previously defined in the Make Color node—for example, white.
Parameter	Type	Description	Before	Thereafter
Detail Level	ChainApprxNone	Saves all points of the contour (very accurate, but memory intensive).
	ChainApproxSimple	Reduces points to a minimal representation (e.g. for a rectangle only 4 corner points).
	ChainApproxTc89L1	A Douglas-Peucker-like method that simplifies contours with fewer points (more accurate but more compact).
	ChainApproxTc89Kcos	Similar to Tc89L1, but uses a different heuristic for simplification (often even smoother approximation).
Retrieval Mode	ListByArea	Returns all contours and sorts them by area (from largest to smallest).
	External	Returns only the outer contours, inner structures are ignored.
	Ccomp	Creates a two-level hierarchy: outer contours and their directly contained holes.
	Tree	Builds a hierarchy of contours (e.g. outer contour and holes in it).

Notes on the parameter set:

Detail Level

In 90% of cases, ChainApproxSimple is the right choice. It's fast, memory-efficient, and provides stable vertices/vertices for most applications.

The other options are useful in the following situations:

ChainApproxNone – when every pixel point is needed (e.g., precise lengths/perimeters on a pixel-by-pixel basis, downstream smoothing/resampling), which, however, consumes memory and time.
ChainApproxTc89L1 / Tc89Kcos – when contours are long or noisy and a compact, smooth representation is required (shape matching, tracking). Kcos is often a bit smoother, while L1 is a bit more angular.

Retrieval Mode

For most applications, External is the recommended setting because it reliably detects the external object contours and hides internal structures.

If the entire hierarchy (e.g., outer contour and all holes/subcontours) needs to be captured, Tree is suitable.
Ccomp is sufficient for a simplified outer/inner relationship.
ListByArea is useful if the largest or smallest contours need to be processed further.

👉 Recommendation: ListByArea or External as standard, Tree if internal structures are important.

Draw Contour

This node allows contours to be displayed directly in the image. This makes it easier to visualize and verify the outlines of previously detected objects or regions.

Important: Before using this node, contours must first be found. To do this, use the Find Contours node and connect its output to the Contours input of this node.

Flow

Parameter set

Parameter	Type	Description
Image	Image → Input	The image in which you want to draw the outlines.
Contours	Polyline Array	A polyline is a continuous line consisting of several connected segments stored in an array that represent the outlines or edges of contours in an image. To do this, a “Find Contours Basic” or “Find Contours Advanced” node must first be used.
Color	Beispiel 1: Green	Choose a color to highlight the outlines. In this example, the outlines are drawn in green. Other colors: Beige Black Blue Green YellowGreen Red White Yellow
	Beispiel 2: Red	Choose a color to highlight the outlines. In this example, the outlines are drawn in red. Other colors: Beige Black Blue Green YellowGreen Red White Yellow
Thickness (Pixel)	Int32	Select the line thickness. A higher number results in a thicker outline. The value must be greater than 1.

Find Pattern

The "Find Pattern" node is used for pattern recognition in images based on contours. This requires a template polyline and the recognized contours as a polyline array. The result provides a score as well as contour data (ContourData), which you then need to process further using the associated "Split" node.

Flow

Parameter set

Parameter	Type	Description
Template Contour	Polyline	A polyline is a continuous line consisting of multiple connected segments stored in an array, representing the outlines or edges of contours in an image.
Contours	Polyline <Array>	A polyline is a continuous line consisting of multiple connected segments stored in an array, representing the outlines or edges of contours in an image.
Match Mode	PatternMatchingMode	Defines the shape-adaptation mode. The modes refer to different calculation methods for the Hu-moment distance. Available options are RelativeDistanceI1, AbsoluteDistanceI2, and WeightedAbsoluteDistanceI3.
Sorting Mode	ContourSortMode	Sets the sorting mode. Available options are score, size, angle, area.
Max Score	Single	Defines the threshold for a match. Contours with higher match values are not returned.
Min Points	Single	Defines the length threshold. Contours with fewer points are not returned.
Return	Contour Data	A contour data array representing the list of contours that match the template contour.

Get Contour By Index

Retrieves contours from a polyline array for a specified index range. Indexing is done by distance from the image origin (0,0), with the nearest contour drawn first. This requires a previous contour search using the "Find Contours Basic" or "Find Contours Advanced" node.

In this way, individual or multiple contours can be specifically selected for further processing.

Flow

Paramerter set

Parameter	Type	Description
Contours	Polyline Array	A polyline is a continuous line consisting of several connected segments stored in an array that represent the outlines or edges of contours in an image. To do this, a “Find Contours Basic” or “Find Contours Advanced” node must be used beforehand.
Min Index	Int32	Minimum index value of the contours to be retrieved. The minimum value is contour length 0.
Max Index	Int32	Maximum index value of the contours to retrieve. The maximum value is contour length - 1.

Get Contour By Area

This node filters contours by area, sorting from largest to smallest. This node is best suited if you're looking for a specific area size or want to filter out all contours. A previous contour search using the "Find Contours Basic" or "Find Contours Advanced" node is required.

Flow

Parameterset

Parameter	Type	Description
Contours	Polyline Array	A polyline is a continuous line consisting of several connected segments stored in an array that represent the outlines or edges of contours in an image. To do this, a “Find Contours Basic” or “Find Contours Advanced” node must be used beforehand.
Min Area Pixel	Int32	Minimum area size of the contours to be included. The minimum value is 0.
Max Area Pixel	Int32	Maximum area size of the contours to be included.

Contours to Region

This node converts contours into a binary image (region) – usable as a mask for further processing. This requires a previous contour search using the "Find Contours Basic" or "Find Contours Advanced" node.

Flow

The return type is a Region. This data type can be converted back to an image format or the image data type using the "Region To Image" node.

Parameter set

Parameter	Type	Description	Example 1	Example 2
Contours	Polyline Array	A polyline is a continuous line consisting of several connected segments stored in an array that represent the outlines or edges of contours in an image. To do this, a “Find Contours Basic” or “Find Contours Advanced” node must be used beforehand.
Region Width (Pixel)	Int32	The width of the binary image, which determines the horizontal size of the mask or region.	Width = 600 Heigth = 600	Width = 1200 Height = 1200
Region Height (Pixel)	Int32	The height of the binary image, which determines the vertical size of the mask or region.	Width = 600 Heigth = 600	Width = 1200 Height = 1200
Mode	Enum: Fill	The contours are filled with white color to create a solid area.
	Enum: Border	Only the edges of the contours are drawn in white.
Return	Region	Binary image processed in black and white.