Gradients
Image Gradient Function
Direction
The gradient points in the direction of most rapid increase in intensity. For an image , the gradient's direction is given by:
Magnitude
The "amount of change" in the gradient is given by its magnitude:
Finite Differences
Discrete Gradient
We want an operator that we can apply to a kernel and use as a correlation (or convolution) filter.
Sobel Operator
The Sobel Operator is a discrete gradient operator that preserves the "neighborliness" of an image that we discussed earlier when talking about the Gaussian blur filter in Blurring Images. It looks like this:
Handling Noise
The images are very noise. To reduce the noise, we apply a smoothing filter.
Derivative Theorem of Convolution
We take advantage of associative of linearity property. This saves us one operation so we can calculate derivatives on the kernal