Liquid Opacity Detection Method Based on Bidimensional Empirical Mode Decomposition

Abstract

According to the problem that liquid turbidity detection is vulnerable to the noise, a novel liquid turbidity detection method based on Bidimensional Empirical Mode Decomposition (BEMD) and Robert operator is proposed. The key part of method is the BEMD algorithm, with which, liquid images can be decomposed to several Intrinsic Mode Functions (IMFs), then we can use Robert operator to detect the edge of each IMF to reconstruct the image edges selectively for highlighting edge details of the liquid and impurity. Experimental results show that the method presented can reduce the influence of random noise on the turbidity detection effectively, and improve the accuracy of turbidity detection.