Fuzzy-PL Transformation based Color Edge Detection

摘要

This paper proposes a new fuzzy multiscale color edge detection approach using power law transformation. The proposed approach involves two stages. In the first stage, color transformation is carried out from RGB color space to HLS color space. The HLS color space posses lightness information much than the RGB color space. Hence, the second stage computes fuzzy symmetry color spread index for each color channels in HLS using multiscale analysis. The edge detection is carried out using fuzzy similarity measure in multiscale analysis. Finally, detected edges are enhanced using power law (PL) transformation that uses tuning parameter (??) for refining the edges. It is experimentally found that proposed approach gives better edge detection at the different levels based on the value of the ?? in the direction of 1 to 0. The results shows that the value of ?? = 0.5 at c = 1 gives the most accurate edges.