Edge detection using fractional order differential operator

Abstract

Edge detection has been widely used in the fields of Image processing, Computer vision, Feature vision n Feature extraction. Edges are detected in an image where the brightness of the image changes abruptly causing discontinuities in image brightness. Many mathematical tools aim at identifying these discontinuities also referred to as edges with in the image. The conventional operators such as Sobel, Robert, Prewitt are Discrete differential operators or often known as integer order differential operators. It was indicated that the edge detection methods operationally are an amalgamation of image smoothing and image differentiation in addition to a post-processing for edge labeling. It was indicated that the edge detection methods operationally are an amalgamation of image smoothing and image differentiation in addition to a post-processing for edge labeling. These integer order differential operators suffer from poor detection accuracy and noise immunity. In the presented work, a fractional order differential operator has been realised in combination of the integer order differential operators. The standard Lena image has been used for performing edge detection operations using both conventional as well as combination of conventional with fractional order differential operators. The improved operator can detect edges with high accuracy, good sharpness and with more detail. The fractional order differential operator has strong capacity to reduce noise than integer order differential operators.