Edge Detection Based on Adaptive Structure Element

The morphological gradient is an attractive option for edge detection which is important in image analysis. But the morphological gradient detector with fixed structure element can hardly suit itself to the diverse property of the edge features, and its performances are constrained. On the spur of this limitation, an improved morphological gradient detector using adaptive structure element is proposed in this paper. Fuzzy topology is introduced into this novel detector for the evaluation of the membership of the pixels and then the shape of the current edge feature, also the shape of the most seemly structure element can be figured out according to this membership. Owing to these efforts, the morphological gradient detector based on the new adaptive structure element wins more powerful capability in the extraction of the variable edge features.