Research on Intelligent Detection Technology of Transparent Liquid based on Style Transfer

In order to solve the problem of transparent liquid perception, this paper proposes a transparent liquid detection method based on style transfer. The generative model based on contrastive learning is used to transform transparent liquid image into colored liquid image, which is equivalent to data e...

Full description

Saved in:
Bibliographic Details
Main Authors: Li, Xianhuai, Zhao, Chenyang, Chen, Yong, Yi, Shihua, Li, Linfeng, Han, Genwei
Format: Conference Proceeding
Language:English
Subjects:
Adaptation models
Cameras
Containers
deep learning
Lighting
liquid detection
Liquids
style transfer
Training
Transforms
Yolov5
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:In order to solve the problem of transparent liquid perception, this paper proposes a transparent liquid detection method based on style transfer. The generative model based on contrastive learning is used to transform transparent liquid image into colored liquid image, which is equivalent to data enhancement of images. After that, the Yolov5 model is used to detect the colored liquid image. The experiment shows that the detection accuracy of this method is up to 99%, and it has good stability. It can accurately detect the transparent liquid in the container, and then help to estimate the area, liquid level height, volume and quality of the transparent liquid.
ISSN:2771-6902
DOI:10.1109/BigDIA56350.2022.9874025