智能系统学报2025，Vol.20Issue(5)：1232-1242,11.DOI:10.11992/tis.202408021
基于电容层析成像传感器的非接触材质识别研究
Electrical capacitance tomography sensor for contactless material recognition
摘要
Abstract
Electrical capacitance tomography(ECT),known for its rapid and nonintrusive characteristics,effectively avoids the optical interference problem commonly encountered in material property identification based on optical ima-ging technologies.However,conventional ECT sensor research has predominantly focused on solving the inverse prob-lem,with limited emphasis on achieving noncontact,nonintrusive material identification through permittivity distribu-tion analysis.To address this gap,this study introduces a planar ECT sensor designed for noncontact material recogni-tion.A material prediction model based on Bayesian-LightGBM is developed,significantly enhancing the predictive performance through Bayesian optimization algorithms.Experimental results demonstrate a high accuracy rate of 95.83％when in contact and 85.28％accuracy within a noncontact range of 20 mm from the sensor.This indicates that robots can precisely acquire material information in the environment in a noncontact and nonintrusive manner,paving the way for new possibilities in the application of robotics technology in complex environments.关键词
电容层析成像/平面电容传感器/传感器建模/非接触识别/材质识别/分类算法/LightGBM/贝叶斯优化Key words
electrical capacitance tomography/planar capacitance sensor/sensor modeling/untouched recognition/ma-terial recognition/classification algorithm/LightGBM/Bayesian optimization分类
信息技术与安全科学引用本文复制引用
许晓丽,郭旭东,郑文栋,刘华平..基于电容层析成像传感器的非接触材质识别研究[J].智能系统学报,2025,20(5):1232-1242,11.基金项目
国家自然科学基金国际合作重点项目(62120106005). （62120106005）
