Supercharge Your Innovation With Domain-Expert AI Agents!

Two-phase flow mixed image segmentation method based on full convolutional neural network

A convolutional neural network and image segmentation technology, applied in neural learning methods, biological neural network models, neural architectures, etc., can solve the problems of blurred gas-liquid boundary, under-segmentation, uneven brightness inside the bubble, etc., to solve the problem of labeling Difficult and precise segmentation

Pending Publication Date: 2022-03-18
KUNMING UNIV OF SCI & TECH
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

The liquid background noise in the gas-liquid image is relatively large, and some bubbles have problems such as uneven internal brightness and blurred gas-liquid boundary. When the traditional segmentation method is used to segment the gas-liquid image, on the one hand, there will be over-segmentation or under-segmentation; On the one hand, it is necessary to determine the corresponding parameters according to different gas-liquid images in order to obtain a better segmentation effect

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
  • Two-phase flow mixed image segmentation method based on full convolutional neural network
  • Two-phase flow mixed image segmentation method based on full convolutional neural network
  • Two-phase flow mixed image segmentation method based on full convolutional neural network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0056] A gas-liquid two-phase image segmentation method based on a fully convolutional neural network specifically includes the following steps:

[0057] Step 1: Prepare the gas-liquid two-phase flow image data set; specifically adopt the scheme of step 101 or step 102:

[0058] Step 101: using the gas-liquid image generated by computer simulation to segment the data set, the data set includes the gas-liquid image and the marked shape and position of the gas-liquid two-phase;

[0059] Step 102: Take a real video of the gas-liquid two-phase flow, and obtain a photo of the gas-liquid two-phase flow from the video to form a gas-liquid two-phase flow picture data set; use the constrained Dirichlet process mixture model (the constrainedDirichlet process mixture model , CDPMM) mark the shape and position of the gas-liquid two-phase for each photo of the gas-liquid two-phase flow.

[0060] Preferably, step 102 includes the following implementation process:

[0061] Step 1021: Deter...

Embodiment 2

[0100] On the basis of the above examples, combined with Figure 1 to Figure 5 , to further illustrate that the gas-liquid two-phase flow bubble image segmentation method of the present invention based on a fully convolutional neural network includes the following steps:

[0101] Step 1: Prepare the gas-liquid two-phase flow image data set; specifically adopt the scheme of step 101 or step 102:

[0102] Step 101: using the gas-liquid image generated by computer simulation to segment the data set, the data set includes the gas-liquid image and the marked shape and position of the gas-liquid two-phase;

[0103] Step 102: Take a real video of the gas-liquid two-phase flow, and obtain a photo of the gas-liquid two-phase flow from the video to form a gas-liquid two-phase flow picture data set; use the constrained Dirichlet process mixture model (the constrainedDirichlet process mixture model , CDPMM) mark the shape and position of the gas-liquid two-phase for each photo of the gas...

Embodiment 3

[0136] On the basis of the above-mentioned embodiments, the present invention evaluates and improves the performance of the FCN model through the following two kinds of data, one is based on a data set generated by computer simulation, and the other is based on a data set obtained from a water model experiment.

[0137] All data analyzes were performed on computers with the following specifications:

[0138] The software environment is based on Window 10, the programming language is Python 3.6.7, and the experiment is completed on the framework of Tensorflow 2.0.0 and Keras 2.3.1. The hardware environment is as follows: the memory is two 3200MHz 8GB memory sticks, the CPU model is i7-10875h, and the GPU model is RTX20608G.

[0139] The present invention adopts the stochastic gradient descent method as the optimization algorithm for training. The hyperparameter values ​​are as follows: learning rate=0.0001, batch size=2, epochs=500.

[0140] For the two-phase flow image segme...

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 a gas-liquid two-phase flow image segmentation method based on a full convolutional neural network (FCN), and belongs to the field of two-phase flow testing. The method comprises the following steps: (1) inputting a two-phase flow mixed image; (2) preprocessing the image; and (3) carrying out semantic segmentation on the two-phase flow image by using an improved FCN network. According to the method, pixel-to-pixel prediction and end-to-end training are carried out on the two-phase flow image, a two-phase flow image segmentation result of any size can be obtained, and in the segmentation process, the problems of repeated storage and convolution calculation caused by using pixel blocks are avoided, so that the calculation efficiency is higher. The method is suitable for segmentation of all images related to two-phase flow and multiphase flow mixing, and the method is good in segmentation effect, high in speed and high in practical value.

Description

technical field [0001] The invention belongs to the field of two-phase flow testing, and in particular relates to a method for segmenting two-phase flow mixed images based on a fully convolutional neural network. Background technique [0002] In the metallurgical industry, the multiphase flow parameters of high-temperature fluids (such as metal melts, high-temperature flue gas, etc.) in metallurgical furnaces are an important factor affecting the metallurgical process. It is not only related to the efficiency of metallurgical reactions, but also determines the The safe service life of the kiln. The mixing characteristics of hot and cold fluids inside the metallurgical furnace are closely related to the heat transfer characteristics caused by direct contact. However, due to the high temperature inside the furnace and the complex environment, it has always been difficult to directly study the mixing effect of the high-temperature molten pool. . In recent years, image process...

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
Patent Type & Authority Applications(China)
IPC IPC(8): G06V10/26G06V10/82G06V10/774G06N3/04G06N3/08
CPCG06N3/084G06N3/088G06N3/045G06F18/2155
Inventor 句媛媛王华肖清泰吴刘仓崔子良杨燕刘冬冬
Owner KUNMING UNIV OF SCI & TECH
Features
  • R&D
  • Intellectual Property
  • Life Sciences
  • Materials
  • Tech Scout
Why Patsnap Eureka
  • Unparalleled Data Quality
  • Higher Quality Content
  • 60% Fewer Hallucinations
Social media
Patsnap Eureka Blog
Learn More

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.

© 2025 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com