Adaptive fuzzy local ternary pattern for mineral flotation froth image edge detection

Abstract

It is difficult to segment blur edge for mineral flotation froth image, an edge detection method based on fuzzy ternary pattern is proposed in this paper. On the basis of binary logic pattern, a kind of fuzzy logic state is introduced. According to the formation mechanism of bubble and the corresponding valley edge characteristics, the local fuzzy ternary pattern algorithm is adopted to describe uncertain intensity relationship between center and neighborhood pixel. Meanwhile, the bidirectional grey scale difference is fuzzed to describe the relationship between edge and non-edge. On this basis, the memberships of neighborhood intensity and bidirectional intensity difference are combined, and the bubble edge membership matrix is built. Thus, the candidate boundary pixels are determined by the joint membership. According to candidate pixels characterized, some non-boundary pixels are removed, and bubble edges are reserved. The experimental results show that the proposed method can effectively detect the bubble edges.