computer vision - When is the gradient direction of an image more useful than the magnitude? -


are there real-world problems or situations directional gradients of image gx , gy more useful magnitude sqrt(gx^2 + gy^2)? intuition, if required, please see matlab image gradient

the gradient information guaranteed perpendicular image (and more grey levels) contours. thus, related geometry of objects in image, not "color". such, information independent actual luminance value of pixel, depends on relative distribution. coined contrast-change-invariant measure.

contrast changes occur in real world applications. take example video surveillance system: during day, sun move on horizon, clouds can fly by, causing light intensity change. system detects gradient orientation changes instead of pixel value of gradient magnitude changes more robust these illumination changes.

more generally, gradient direction, unlike magnitude, closely related vast field of image morphology useful in shape recognition context.


Comments

Popular posts from this blog

basic authentication with http post params android -

vb.net - Virtual Keyboard commands -

css - Firefox for ubuntu renders wrong colors -