The proposed approach to the detection of diffuse edges is based on three ordered operations: noise reduction, gradient generation, and edge formalization. In the first stage, an efficient two-stage median filtering scheme is introduced for noise reduction. Next, a generalized second-order gradient generation algorithm based on an efficient averaging algorithm is applied. Finally, an edge strength normalization and thresholding scheme is introduced for the edge formalization stage to eliminate the breaking up of edge contours. Simulation results demonstrating the superior performance of the proposed approach are included.