Edge Detection Using Simple Image Arithmetic

The process of edge detection is commonly used as a fundamental pre-processing step in image processing, particularly in the areas of feature detection and feature extraction. Conventional edge detection methods use a variety of mathematical methods to detect sharp changes in image brightness. In this paper however, another method of detecting edges in images is proposed using simple image arithmetic. The proposed method involves converting the image into grayscale, sharpening a copy, smoothening a copy and then performing the image arithmetic followed by binarization. This results in the formation of a logical image with the edges of the original image highlighted including the thick edges and bold letters. This method of detecting edges is simpler and can be easily changed according to the user's wishes and is more useful in processes like letter extraction. The proposed method can be easily incorporated into any application that requires edge detection without much in-depth knowledge about image processing. It can also be a valuable image preprocessing technique. Edge detection is also the key step in various real world applications like handwriting recognition, biometrics such as fingerprint matching and remote sensing etc.