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

Molten iron slagging-off detection method based on lightweight convolutional neural network

A technology of convolutional neural network and detection method, which is applied in the field of molten iron slag detection based on lightweight convolutional neural network, can solve the problems of unsatisfactory recognition accuracy and precision, and avoid manual participation in harsh and complex environments. Anti-jamming ability and effect of improving accuracy

Active Publication Date: 2019-12-03
UNIV OF SCI & TECH BEIJING
View PDF7 Cites 6 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0007] The technical problem to be solved by the present invention is to provide a method for detecting molten iron slag based on a lightweight convolutional neural network, which requires only The video stream image of the molten iron ladle slag removal collected by the monitoring camera system is input to the embedded device for processing, and the remote monitoring of the ladle slag removal status can be realized

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
  • Molten iron slagging-off detection method based on lightweight convolutional neural network
  • Molten iron slagging-off detection method based on lightweight convolutional neural network
  • Molten iron slagging-off detection method based on lightweight convolutional neural network

Examples

Experimental program
Comparison scheme
Effect test

no. 1 example

[0038] This embodiment provides a method for detecting molten iron slag based on a lightweight convolutional neural network. First, a video is collected from the actual industrial site of molten iron slag extraction, and enough pictures are extracted from it to make a data set, including 300 original images. Then, under the guidance of experienced workers, manually label these pictures to obtain the corresponding label data. Then, according to the principle of deep learning to classify each pixel of the image, the data set is trained using the lightweight network model BiseNet (Bilateral Segmentation Network for Real-time Semantic Segmentation).

[0039] BiseNet contains two paths, Spatial Path and Context Path, among which Spatial Path can preserve the spatial scale of the original input image and encode rich spatial information. Spatial Path consists of three layers, each layer contains a convolution with a stride of 2, followed by batch normalization and ReLU activation fun...

no. 2 example

[0052] This embodiment provides a method for detecting hot metal slag scraping based on a lightweight convolutional neural network, including:

[0053] 1. Collect a 3200s video of molten iron slag removal from the actual industrial site of molten iron slag removal through an industrial camera; among them, the video resolution is 1920x1080, and the frame rate is 25 frames per second;

[0054] 2. Take one frame every 8 seconds from the collected molten iron slag removal video, and get 400 original images; divide the obtained original images into training data set and test data set according to the ratio of 3:1; among them, the training data set Under the guidance of experienced people, use PhotoShop to manually annotate the 300 pictures at the pixel level. These 300 pictures are used as the training data set, and the remaining 100 pictures are used as the test data set;

[0055] 3. During training, through random up-down flip, random left-right flip, and random size cropping, da...

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 molten iron slagging-off detection method based on a lightweight convolutional neural network. The molten iron slagging-off detection method comprises the following steps: collecting a molten iron slagging-off video with a preset duration from an industrial site of molten iron slagging-off; extracting a preset number of pictures from the collected molten iron slagging-offvideo, and constructing a training data set; labeling each picture in the constructed training data set to obtain corresponding label data; and training the training data set by using the lightweightnetwork model, and detecting the molten iron slagging-off pictures acquired in real time by using the trained lightweight network model in a framing manner to monitor the molten iron slagging-off condition in real time. The method provided by the invention has better anti-interference capability, improves the precision of molten iron and desulfurization slag detection in molten iron slagging-off,can effectively detect molten iron and desulfurization slag, and can also effectively identify the inner ladle wall of molten iron and a slagging-off device at the same time. The molten iron slagging-off process can be stably detected, meanwhile, the real-time performance can meet the actual monitoring requirement, and great application value is achieved in the production process.

Description

technical field [0001] The invention relates to the technical field of deep learning, in particular to a detection method for molten iron slag removal based on a lightweight convolutional neural network. Background technique [0002] In the steelmaking process, if the molten iron is to be sent to the converter for steelmaking, it must undergo one-step desulfurization treatment. However, if only the desulfurization treatment is performed, the sulfur content in the molten iron cannot be completely removed, because there is a large amount of desulfurization slag on the surface of the molten iron after desulfurization treatment. If this layer of desulfurization slag cannot be disposed of in time, the sulfur element in the slag will be reduced by oxygen. . The sulfur content of molten steel is an important parameter in the steelmaking process. If the slag removal is not clean enough before the molten iron is poured into the converter from the ladle, the sulfur content of the mol...

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): G06K9/00G06K9/62G06N3/04G06N3/08G06T7/00
CPCG06N3/084G06T7/0008G06T2207/10016G06T2207/30116G06V20/41G06N3/045G06F18/24G06F18/214
Inventor 李江昀郑俊锋俞洪
Owner UNIV OF SCI & TECH BEIJING
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