Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

A Semantic Segmentation Method of Solution Crystallization Process Image Based on Improved Unet Model

A solution crystallization and process image technology, applied in image analysis, image enhancement, image data processing, etc., to reduce aliasing effect, improve accuracy, and reduce information loss

Active Publication Date: 2022-04-01
BEIJING INSTITUTE OF PETROCHEMICAL TECHNOLOGY
View PDF14 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, inherent flaws in feature fusion prevent feature pyramid networks from further aggregating more discriminative features

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
  • A Semantic Segmentation Method of Solution Crystallization Process Image Based on Improved Unet Model
  • A Semantic Segmentation Method of Solution Crystallization Process Image Based on Improved Unet Model
  • A Semantic Segmentation Method of Solution Crystallization Process Image Based on Improved Unet Model

Examples

Experimental program
Comparison scheme
Effect test

Embodiment

[0037] A method for image semantic segmentation of solution crystallization process based on the improved Unet model, the structural block diagram is as follows figure 1 shown, including the following steps:

[0038] S1: Construct a crystallographic semantic segmentation dataset based on crystallographic image data, and preprocess the images in the dataset; preprocessing includes image cropping, correction, denoising and enhancement.

[0039] S2: The preprocessed picture is input into the improved Unet model; the improved Unet model is to add a channel-enhanced feature pyramid network between the downsampling part and the upsampling part of the Unet typical model, such as figure 2 shown.

[0040] Among them, the channel enhanced feature pyramid network includes a context enhancement module and a channel attention module:

[0041] S2.1: The structure of the context enhancement module is as follows image 3 shown, including:

[0042] Obtain the output of the first sampling ...

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 provides a solution crystallization process image semantic segmentation method based on the improved Unet model, comprising: constructing a crystal semantic segmentation data set according to the crystallization image data; inputting the preprocessed picture into the improved Unet model; The channel-enhanced feature pyramid network is added between the down-sampling part and the up-sampling part of the typical model; the channel-enhanced feature pyramid network adds an up-down enhancement module and a channel attention module on the basis of the feature pyramid network; the improved Unet model is trained through the dataset , evaluate the segmentation effect of the improved Unet model, and obtain the improved Unet model after training; input the crystal image data to be segmented into the improved Unet model after training to obtain the segmentation result. Compared with Unet semantic segmentation models based on other feature pyramid networks, the improved Unet model based on channel enhancement pyramid in the present invention can effectively improve the accuracy of measurement without adding additional complex operations.

Description

technical field [0001] The invention relates to the technical field of image semantic segmentation, in particular to a method for image semantic segmentation in a solution crystallization process based on an improved Unet model. Background technique [0002] Crystallization is the operation of separating a substance from a solution in a crystalline state. It can solve many problems that cannot be solved by operations such as unit distillation, extraction, and adsorption, and is widely used in the process of separating new products. In the crystallization process, the focus and purpose of process detection is to describe and control information such as crystal morphology and crystal size. Usually, measuring crystal morphology and size distribution during crystallization is one of the important means to control and optimize to obtain desired product quality and production efficiency. At present, process analysis tools have begun to be applied to the crystallization process, s...

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 Patents(China)
IPC IPC(8): G06T7/10G06V10/40G06N3/04G06N3/08
CPCG06T7/10G06N3/08G06T2207/20016G06N3/045
Inventor 朱亚东洋李心超何运良李婧莹高梦楠张立立李晶崔宁
Owner BEIJING INSTITUTE OF PETROCHEMICAL TECHNOLOGY
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

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

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com