An Improved Discrete Wavelet Transform based Edge Detection Algorithm for Noisy Images

In this paper, an improved DWT based edge detection algorithm for noisy images has been proposed. The proposed edge detection method works efficiently on images influenced by noise and is able to differentiate between noise and real edges, thus detecting the actual edges. In the proposed algorithm the DWT edge detector separates detail coefficients at the time of decomposition, thus it has to eliminate the residual noise and insignificant edges from the approximation coefficients which makes the algorithm faster and give good results. Classical edge detectors like Roberts, Sobel, Prewitt, Laplacian and Canny fail to detect edges in noisy images. To evaluate the vulnerability of the proposed edge detector to noise, qualitative and quantitative analysis is made and the PSNR of proposed edge detector on image with Gaussian noise is compared with Canny and wavelet transform based edge detector and it is found that proposed method performs better.