Morphological image processing is a technique for modifying the pixels in an image. In the case of a grayscale image the pixels are identified by the binary values of 0 and 1, and the process is conducted using either sophisticated image processing algorithms or less mathematically complicated operations. These include erosion and dilation as well as opening and closing. The purpose of morphological image processing is to remove unwanted artifacts from an image or improve its clarity. It is used in such applications as fingerprint processing, viewing pictures from space telescopes, and analyzing medical scans.
An object in an image is represented by a specific set of pixels, called object pixels. Background pixels are represented separately and are white. The operation of erosion converts pixels associated with the object’s boundary to pixels in the background, while with dilation, the bordering background pixels are changed to ones associated with the object. Objects become smaller during the erosion process, and enlarge or even merge during dilation.
The two operations can be combined in morphological image processing, so an image can be edited by performing erosion and then dilation, which results in opening. Filaments and isolated pixels can be removed from the object in this way to smooth out the image. Background pixels can be filtered using the closing operation, which can remove holes and pixels that are already known to not be in the right place. Another morphological image processing technique is called skeletonization, during which extra pixels can be removed to form single lines. It is often used to process fingerprints.
Image processing applications use a few rules to change the image visualization, or utilize set theory, a mathematical concept that is often more sophisticated than required. In changing a pixel from object to background, the program only considers the pixels associated with the object. It also concentrates on the edge regions, so neighboring background pixels are analyzed before the black pixels can be changed. If an object pixel is to be changed, then more than one similar pixel must be bordering it as removing pixels at the end of a line can distort the image.
A morphological image processing program is based on the idea of keeping objects whole. If taking away a pixel will break up a single object, then the program will not remove it. Image processing technology consists of various software programs that can allow for playback of image changes, rewind of changes to get back to preferred configurations, and analysis on how specific changes affect parts of an image.