Unlock instant, AI-driven research and patent intelligence for your innovation.

ECT two-dimensional image reconstruction method based on dilated convolutional neural network

A convolutional neural network and two-dimensional image technology, applied in the field of capacitance tomography, can solve the problem of low image reconstruction accuracy, achieve the effect of increasing the number of measured capacitances and increasing the amount of information

Pending Publication Date: 2022-05-03
HARBIN UNIV OF SCI & TECH
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0003] The purpose of the present invention is to propose a two-dimensional image reconstruction method for ECT based on a hollow convolutional neural network to improve the image reconstruction of the ECT system in view of the underdetermined problem in the ECT system and the low accuracy of image reconstruction caused by soft field characteristics the accuracy of

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • ECT two-dimensional image reconstruction method based on dilated convolutional neural network
  • ECT two-dimensional image reconstruction method based on dilated convolutional neural network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0012] The sensor of the ECT system is modeled and designed by ANSYS18.0 finite element analysis software, and the ECT positive problem is simulated by the finite element method. The inner diameter of the 24-electrode ECT sensor pipe is 76.8mm, the outer diameter of the pipe is 81.8mm, and the shielding cover is 106.8mm.

[0013] For the 24-electrode ECT system, use ANSYS18.0 for grid division, and then use the APDL language CMATRIX macro definition command to solve the capacitance value. For the 24-electrode ECT system, 276 capacitance values ​​can be measured. For the 24-electrode ECT system, the capacitive sensor model is The center of the pipe cross-section is symmetrical to the center, and the 24 electrodes are evenly arranged on the outside of the pipe, and the parameters are exactly the same, so that it is guaranteed that the 24 electrodes have rotation symmetry and can be interchanged.

[0014] The line connecting the center of the pipe section to the center of the capa...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention discloses an ECT two-dimensional image reconstruction method based on a dilated convolutional neural network, which mainly measures a phase flow in a closed pipeline, obtains a capacitance value of a medium in which the measured phase flow flows in the pipeline through a capacitance sensor, and mainly adopts ANSYS18.0 software to carry out model design of an ECT sensor. And detecting the capacitance value of the phase flow in the pipeline. According to a traditional ECT technology image reconstruction algorithm, a pixel gray value of a corresponding position of an image is calculated by using sensitivity distribution representing a capacitance measurement value and a dielectric constant distribution relation of a measured area. However, the sensitivity distribution is easily influenced by the dielectric constant distribution of the measured multiphase flow, the distribution is non-uniform in the measured area, a good effect is obtained on the dilated convolutional neural network, and a nonlinear activation function is added to the convolutional layer and the full-link layer of the convolutional neural network, so that the method can be well applied to the nonlinear relationship between the capacitance value and the measured area. And two-dimensional pipeline image reconstruction can be highly realized.

Description

technical field [0001] The invention belongs to the technical field of electrical capacitance tomography, and relates to a two-dimensional image reconstruction method based on a hollow convolutional neural network ECT. Background technique [0002] The sensor design of the ECT system is mainly based on a 24-electrode sensor proposed in the original design of the 8-electrode and 12-electrode sensors. The underdetermined problem of the ECT system is mainly because the measured capacitance value is smaller than the pixel points of the image reconstruction. , the capacitance value measured by 12 electrodes is 66 capacitances, and the sensor setting of 24 electrodes can measure 276 capacitance values, which can minimize the gap between the capacitance value and the pixel point of image reconstruction, thereby ensuring the accuracy of image reconstruction, data The acquisition of the set is mainly obtained by using C++ and APDL language mixed programming, using ANSYS18.0 script la...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
IPC IPC(8): G06F30/23G06F30/27G06N3/04G06N3/08G06T11/00
CPCG06F30/23G06F30/27G06T11/003G06N3/08G06N3/045
Inventor 韩文双李岩姚文杰
Owner HARBIN UNIV OF SCI & TECH