Multi Classification ERT Flow Pattern Recognition Method Based on Deep Learning

Abstract Electrical resistance tomography (ERT) is the frontier technology of modern industrial detection, in which flow pattern an important index two-phase detection. Affected by many factors, ERT recognition difficult. In this paper, method based on deep learning designed order to obtain real situation pipeline practical application. The original measured voltage transformed from one-dimensional data information two-dimensional dot matrix pseudo image coding method. According characteristics, patterns are divided into 27 categories, and then databases with different scales established time domain frequency domain. Convolutional neural network used construct model learning, experiments verify its performance. results show that average accuracy each algorithm can reach 98.74%, 14 types 100%. This achieve high-precision task.