Back to Adaptive Vision Library Lite website

You are here: Start » Image Local Transforms » GradientMagnitudeImage



Measures the strength of gradient at each pixel location with Sobel or Prewitt operator.


void avl::GradientMagnitudeImage
	const avl::Image& inImage,
	atl::Optional<const avl::Region&> inRoi,
	avl::GradientMaskOperator::Type inOperator,
	avl::MagnitudeMeasure::Type inMeasure,
	const int inScale,
	avl::Image& outValueImage


Name Type Range Default Description
inImage const Image& Input image
inRoi Optional<const Region&> NIL Range of output pixels to be computed
inOperator GradientMaskOperator::Type Defines how the gradient is computed
inMeasure MagnitudeMeasure::Type Hypot Defines how the gradient magnitude is computed
inScale const int 1 - 16 1 Scales the resulting gradient magnitudes
outValueImage Image& Gradient magnitudes of the image


The operation computes the magnitude of the intensity change at each pixel of the inImage. Firstly the selected inOperator is used to obtain two-dimensional gradient vector at each pixel. Then the magnitudes of the vectors are estimated using the inMeasure method.

Specified by inMeasure method computes magnitude (A) from horizontal gradient component (x) and vertical gradient component (y) using one of following formulas:
\[A_{Horizontal}=|x|\] \[A_{Vertical}=|y|\] \[A_{Average}=\frac{|x|+|y|}{2}\] \[A_{Sum}=|x|+|y|\] \[A_{Maximum}=Max(|x|,|y|)\] \[A_{Hypot}=\sqrt{x^{2}+y^{2} }\]

The magnitudes are multiplied by inScale factor and saturated if they exceed the greatest value of their type.


GradientMagnitudeImage performed on the sample image with inOperator = Sobel, inMeasure = Hypot.

Hardware Acceleration

This operation is optimized for SSSE3 technology for pixels of type: UINT8.

This operation is optimized for AVX2 technology for pixels of type: UINT8.

This operation is optimized for NEON technology for pixels of type: UINT8.

This operation supports automatic parallelization for multicore and multiprocessor systems.