In this chapter, we shall continue our discussion on image enhancement, which is the problem of improving the appearance or usefulness of an image. We shall concentrate mainly on spatial filtering techniques to compute image gradients/derivatives, and how these techniques can be used for edge detection in an image. First, we shall start with the basic concepts of image gradients using the first order (partial) derivatives, how to compute the discrete derivatives, and then discuss the second order Derivative/Laplacian. We shall see how they can be used to find edges in an image. Next, we shall discuss a few ways to sharpen/unsharp mask an image using the Python image processing libraries PIL, the filter module of scikit-image, and the ndimage module of SciPy. Next, we shall see how to use different filters (sobel, canny, LoG, and so on) and convolve them with the image to detect edges in an image. Finally, we shall discuss how to compute Gaussian/Laplacian image pyramids (with scikit-image) and use the image pyramids to blend two images smoothly. The topics to be covered in this chapter are as follows:

• Image Derivatives—Gradient, Laplacian
• Sharpening and unsharp masking (with PIL, scikit-image, SciPy ndimage)
• Edge detection using derivatives and filters (Sobel, Canny, LOG, DOG, and so on with PIL, scikit-image)
• Image pyramids (Gaussian and Laplacian)—Blending images (with scikit-image)