Lightweight bubble detection algorithm for quartz 
crucibles based on improved YOLOv5

ZHAO Qian; ZHENG Xuan; ZHAO Man

Chinese

您当前的位置：

首页 >

文章列表页 >

Lightweight bubble detection algorithm for quartz crucibles based on improved YOLOv5

Calculation & Test | 更新时间：2025-12-30

- Lightweight bubble detection algorithm for quartz crucibles based on improved YOLOv5
- Vol. 44, Issue 10, Pages: 125-128(2025)
- 作者机构：
  
  1. 西安科技大学通信与信息工程学院,陕西,西安,710054
  2. 西安地山视聚科技有限公司,陕西,西安,712044
- 作者简介：
- 基金信息：
- DOI：
  CLC： TP391;TN307
- Published：2025
- 稿件说明：
移动端阅览
ZHAO Qian, ZHENG Xuan, ZHAO Man. Lightweight bubble detection algorithm for quartz crucibles based on improved YOLOv5[J]. 2025, 44(10): 125-128.
DOI：

ZHAO Qian, ZHENG Xuan, ZHAO Man. Lightweight bubble detection algorithm for quartz crucibles based on improved YOLOv5[J]. 2025, 44(10): 125-128. DOI：

摘要

针对现有的石英坩埚气泡检测准确度低和速度慢的问题，提出一种基于YOLOv5n的轻量化气泡检测算法。首先，在主干网络采用C3Ghost模块替换C3模块，然后使用GSConv替换颈部网络的Conv，同时使用VoVGSCSP模块替换颈部网络的C3。实验结果表明，在自建的石英坩埚透明层气泡数据集上，改进后算法mAP为98.55%，计算量为2.7G，权重大小为2.54MB，帧率为137fps，后期方便部署到嵌入式设备。

Abstract

To address the issues of low accuracy and slow speed in existing quartz crucible bubble detection

a lightweight bubble detection algorithm based on YOLOv5n is proposed.Firstly

the C3 module in the backbone network is replaced with the C3Ghost module.Then

GSConv is used to replace the Conv in the neck network

and the VoVGSCSP module is used to replace the C3 in the neck network.Experimental results show that on a self-built dataset of transparent layer bubbles in quartz crucibles

the improved algorithm achieves mAP of 98.55%

a calculated amount of 2.7G

a weight size of 2.54MB

and a frame rate of 137fps

making it suitable for deployment on embedded devices in the future.

关键词

Keywords

references

Views

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

Self-supervised guided filtering for lightweight depth estimation  in complex visual systems

Automatic recognition system for liquid dripping based on inter-frame difference calculation

Research on improved median filtering algorithm fused with gradient information

Indoor localization algorithm based on image processing and RSSI signal

Research on restoration technology of motion blur image

Related Author

ZHANG Yawei

LI Xiaoguang

XU Guanhua

GU Tianhao

TANG Jun

YUAN Ni

ZHAO Boyang

ZHAO Jinke

Related Institution

School of Automation,Qingdao University

Institute of Intelligent Unmanned Systems,Qingdao University

School of Electronic Engineering Xi’an Shiyou University

School of Mechanical Engineering,Nanjing University of Science and Technology

Branch Factory,Heilongjiang North Tool Limited Company

⁰