Hybrid Edge Detector

Abstract

In this paper a hybrid edge detection algorithm has been envisaged. Real Images are often corrupted by noise from various sources. The classical edge detectors fail to give adequate intelligence of edges for images with depleted signal to noise ratio (SNR). A major dilemma in edge-detection for noisy images is the choice of optimum threshold which lacks generality. The dominance of noise in image reflects false edges in spatial domain. Addressing this dilemma, this paper presents a methodology and a framework in which smoothing effects of wavelet filters have been exploited prior to classical operators. Experimental results indicate that the hybrid edge detector gives qualitative and quantitative results comparable to Canny (1986), while it is simpler to implement.