Quasi-triangular norms to electrostatic edge detection approach

Recent years have witnessed significant advancements in image processing, particularly in edge detection techniques, which play a crucial role in various computer vision applications. However, accurately detecting edges while preserving fine details and ensuring robustness to noise remains a challenging task. To address these limitations, this paper introduces an effective technique to perform edge detection by considering the referenced electrostatic edge detection (EED) method by B. Bouda et al. and applying quasi triangular norms on it. This novel method enhances the capabilities of the EED approach and addresses several of its limitations such as thick edges, missing edge points, and a lack of robustness to noise. We begin by addressing key aspects of the original EED method, namely the consideration of null-charge pixels, the noise contamination, and the homogeneous distribution of charge pixels in the image, along with several proposed improvements. Next, we replace the scalar product with quasi triangular norms in the computation of electrostatic forces. Finally, we construct a set of edge points by the resultant forces obtained from applying quasi t-norms to EED approach at each pixel. Experimental results demonstrate that the proposed method significantly enhances edge detection quality in terms of accuracy and robustness.