Coagulation Detection Method Based on the Image Information Entropy from a New Microscopic Dew-point Hygrometer

Abstract

It is difficult to accurately identify coagulation timing and phase of coagulation for dew-point hygrometer. In this paper, an imaging dew-point device is developed and a method is presented to extract textural feature of the information entropy of the mirror image of dew-point instrument according to image texture characteristics of different types of mirror. The method of detection and identification of condensation of dew-point instrument based on texture was designed. Mirror images are first extracted the textural feature using the gray scale symbiotic matrix. The information entropy of the image is extracted, and then the characteristics of entropy of different types of images were analyzed. The entropy value threshold of different types of images was established and tested. It is shown that we are able to achieve an average detection accuracy rate of 97.8% for different types of mirrors. It is a great solution combined with the continuous observation method of the detection and identification of dew-point condensation.