Particle reconstruction method and device in three-dimensional flow field, electronic device and storage medium

A three-dimensional flow field and particle technology, applied in the field of flow field, can solve the problem of insufficient accuracy of the three-dimensional particle field, and achieve the effect of high reconstruction accuracy, high accuracy and small amount of calculation.

Pending Publication Date: 2019-12-27
ZHEJIANG UNIV
View PDF2 Cites 1 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, the three-dimensional particle field obtained

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
  • Particle reconstruction method and device in three-dimensional flow field, electronic device and storage medium
  • Particle reconstruction method and device in three-dimensional flow field, electronic device and storage medium
  • Particle reconstruction method and device in three-dimensional flow field, electronic device and storage medium

Examples

Experimental program
Comparison scheme
Effect test

example 1

[0121] Such as Figure 8 As shown, this example provides a particle reconstruction method in a three-dimensional flow field, including:

[0122] The first step is the initialization of the three-dimensional particle field. After reading in the image captured by the camera, use an initialization algorithm to obtain the initial three-dimensional matrix. There are many initialization algorithms, including but not limited to: MLOS algorithm, MFG algorithm or constant initialization algorithm. The obtained initial three-dimensional matrix can be expressed as a three-dimensional matrix. The initial three-dimensional matrix does not provide accurate information on the position and shape of the particles, but only projects the particles captured on the two-dimensional image in the direction of the line of sight to obtain a three-dimensional matrix. The initial three-dimensional matrix is ​​used as the input of the three-dimensional reconstruction algorithm to carry out the three-dim...

example 2

[0131] The image captured by the camera is read in, and the initialization method is used to obtain the initial three-dimensional matrix. The method used in this embodiment is the MLOS algorithm. The size of the initial three-dimensional matrix of particles after initialization is 256×256×128, see the cross-section image 3 . As shown in the figure, there is no obvious particle shape and position information in the initial three-dimensional matrix

[0132] The matrix block (i.e. block processing) of the initial three-dimensional matrix can be as follows:

[0133] The block processing of the initial three-dimensional matrix may include: dividing the initial three-dimensional matrix into small blocks of 64×64×32. In this embodiment, the field after the initialization of the MLOS algorithm is divided into 64 blocks (4×4×4) according to the method of adjacent block division (ie, block by block). For the section image after block, see Figure 4 .

[0134] Applying the convolut...

example 3

[0139] This example also provides a neural network that can be used for 3D reconstruction of particles based on the initial 3D matrix. The structure of the neural network can be as follows Figure 9 As shown, including: input layer; middle layer; output layer.

[0140] The neural network may include predetermined network layers, for example, 12 network layers, the first network layer is an input layer, the last network layer is an output layer, and the remaining 10 network layers are intermediate layers.

[0141] The network layers are connected layer by layer to form the network architecture.

[0142] Figure 9 Shown is a 3D convolutional neural network with 12 network layers. The size of the input layer and the output layer can be both 64×64×32×1, the size of the convolution kernel can be 3×3×3, the activation function of the last layer is the Sigmoid function, and the activation function of the other network layers can be used Linear Correction Unit (Relu) function. The...

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 particle reconstruction method and device in a three-dimensional flow field, an electronic device and a storage medium. The particle reconstruction method in the three-dimensional flow field comprises the following steps: acquiring a two-dimensional image formed by collecting particles in the three-dimensional flow field; obtaining an initial three-dimensional matrix based on an initialization algorithm according to the two-dimensional image; inputting the initial three-dimensional matrix into a neural network; and obtaining a three-dimensional particle field based onthe output of the neural network.

Description

technical field [0001] The invention relates to the field of flow field technology, in particular to a particle reconstruction method and device in a three-dimensional flow field, electronic equipment and a storage medium. Background technique [0002] In actual industrial requirements, a large number of scenarios require the velocity measurement of the fluid in the three-dimensional flow field. The velocity measurement method of the two-dimensional flow field is to use the camera to shoot the particles in the velocity field at different moments, and then calculate the velocity field through the subsequent algorithm. The particles in the two-dimensional velocity field can be clearly captured by the camera, but the camera cannot capture three-dimensional images. The solution to the measurement of the three-dimensional velocity field is generally to shoot the particles in the velocity field from different angles through multiple cameras set up in different positions, and then...

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): G06T17/00G06T15/00
CPCG06T17/00G06T15/005G06T2207/20084
Inventor 高琪李其杰魏润杰
Owner ZHEJIANG UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

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

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap