Convolution is a basic operation in image processing, where we assign to each pixel x a weighted sum of all pixel values f(x) in a window centered in x. The weights given to the pixel values in the window will define what the convolution does to the image. Examples include image smoothing, image sharpening, image gradients.
Convolutions are the basis for many important methods in image analysis, including Wavelet transform and Convolutional Neural Networks.