机电工程技术2024，Vol.53Issue(2)：44-46,3.DOI:10.3969/j.issn.1009-9492.2024.02.009
透明容器液面机器视觉检测方法
Machine Vision Method for Liquid Level Detection in Transparent Containers
摘要
Abstract
The key to detecting the volume of liquid in a container using machine vision is liquid level image recognition.Based on the imaging features of the target,machine vision detection methods for transparent container liquid level based on liquid level features and reference features are studied.Through image examples and imaging mechanism analysis,it is found that the liquid level detection method using grayscale and edge features of the liquid level is greatly affected by the shape and imaging quality of the liquid level.However,the liquid level detection method using external reference refraction imaging with separation bands is also prone to detection failure when industrial cameras view the liquid level horizontally,and the top-down search method has a problem of large detection errors.In order to improve detection stability,a new method for liquid level detection based on external reference imaging features is proposed.The method is based on the rule that transparent containers affect the imaging of reference objects as the detection principle.By detecting the inter line gradient of the reference object and utilizing the feature of the maximum gradient at the liquid level,the liquid level is searched and recognized from bottom to top.The new detection method not only enhances the robustness of liquid level recognition,but also improves the accuracy of liquid level detection.关键词
机器视觉/液面检测/图像识别/灰度/梯度Key words
machine vision/liquid level detection/image recognition/grayscale/gradient分类
信息技术与安全科学引用本文复制引用
孙晓辉,朱洪雷,蔡永洪..透明容器液面机器视觉检测方法[J].机电工程技术,2024,53(2):44-46,3.基金项目
2022年度广州市高等教育教学质量与教学改革工程项目(2022JXTD027) （2022JXTD027）
2020年广东省高职院校机器人与智能装备技术产教融合创新平台项目(2020CJPT026) （2020CJPT026）
广东省市场监督管理局科技项目(2020CJ02) （2020CJ02）
