An Adaptive Threshold Segmentation Algorithm for Impurity in Liquid Automatic Detecting System

A novel algorithm for detecting impurity-one kind of small target in liquid by analyzing sequences of grey scale images is described, which is able to detect impurity with full automatic. It is based on the basic theories of fuzzy logic and higher-order statistics and is divided into three main stages: moving targets extraction, adaptive thresholds selection and target tracking. Each of them is explained in detail in this paper. In the first stage, a new method of extracting information of moving targets in visible image sequences is proposed, which is based on the fourth-order moment of the summation of inter-frame difference with 5 continuous frames. In order to changing the HOS thresholds adaptively, in the second stage, a fuzzy logic controller is adopted to adjust the threshold adaptively according to the variance and mean of the summation difference image. The experimental and factual testing results show that the proposed algorithm can meet the demands of impurity in medicinal liquid in real-time detecting, and that it is a practicable and effective image segmentation method.