基于视觉检测的非接触式膏体浓度识别方法
Hits:
DOI number:10.11817/j.issn.1672-7207.2023.05.027 Journal:中南大学学报 Abstract:为解决膏体料浆对于接触式浓度计的损坏以及核子浓度计对施工人员的辐射等问题，提出一种基于视觉检测的非接触式膏体浓度识别方法，以膏体图像数据集为基础，采用卷积神经网络模型提取膏体深度特征，设计多模型全连接层融合算法、伽马矫正数据增广策略，对膏体浓度监测准确率进行分析。研究结果表明：手动搅拌膏体数据集的准确率为88.79%，在自动搅拌膏体数据集上获得91.42%的准确率，可有效提高浓度识别准确性，避免浓度计损坏或辐射危险 Co-author:周佳城,Banxiaojuan,袁兆麟,Ruanzhuen First Author:Maboyuan Indexed by:Journal paper Discipline:Engineering Document Type:J Volume:54 Issue:05 Page Number:1-12 Translation or Not:no Date of Publication:2023-01-01 Included Journals:EI -
|
 ZipCode:9afbadb1bdcd21f2605641205359fae42a33050a01aa43854cbd3c2c2fba7e51e00e98dd4ebb298a9e075bea70de2dc35b240e3c4868702506040a6d18b342786082da32963b6185d3a8a00d9746a41e2dea73822dae12c86efb51012eb15fb38a03b0c3dd242fbea002f942170f137c7639f1a296486404caf634982f5aab08 PostalAddress:73941b85a48ff0bf4c22e5699874367d8b3786bd0242d1955eb29c74dd5f84ce7bf3e296bb25725e9e35fe490f9d6e9838adf246c1214740869a6f13c891e46a6efc1d63db8964d21c6b27285809b804d9af9abf250775985bfa2ef027b0b51a1195148ed4da9b381910c290446fc482d5830d594b9c1759ab60eba74ce04413 Email:7a0d2b957c40b9e2da0a01b15583d983a26031fa23f266a6def5772ce24922830e13c337f928d3b9f441e3cd7f43b39869da167205118ff405bb767cf06524dca6ce03cda1034eaff073d2f37ab3cfbb3b7e46ee05f4b82f779ec434c2854fe4f4e46d111aa9b1e23b0f422e98fcf83422a59929317e2a497fb8ef8dbb37ac33
|
MOBILE Version
