A machine vision system for the calibration of digital thermometers
Identifieur interne : 000503 ( Istex/Corpus ); précédent : 000502; suivant : 000504A machine vision system for the calibration of digital thermometers
Auteurs : Esteban Vzquez-Fernndez ; Angel Dacal-Nieto ; Higinio Gonzlez-Jorge ; Fernando Martn ; Arno Formella ; Victor Alvarez-ValadoSource :
- Measurement Science and Technology [ 0957-0233 ] ; 2009.
Abstract
Automation is a key point in many industrial tasks such as calibration and metrology. In this context, machine vision has shown to be a useful tool for automation support, especially when there is no other option available. A system for the calibration of portable measurement devices has been developed. The system uses machine vision to obtain the numerical values shown by displays. A new approach based on human perception of digits, which works in parallel with other more classical classifiers, has been created. The results show the benefits of the system in terms of its usability and robustness, obtaining a success rate higher than 99 in display recognition. The system saves time and effort, and offers the possibility of scheduling calibration tasks without excessive attention by the laboratory technicians.
Url:
DOI: 10.1088/0957-0233/20/6/065106
Links to Exploration step
ISTEX:EBF303D3EB21A4EED751C98AB5B05281A086F542Le document en format XML
<record><TEI wicri:istexFullTextTei="biblStruct"><teiHeader><fileDesc><titleStmt><title>A machine vision system for the calibration of digital thermometers</title><author><name sortKey="Vzquez Fernndez, Esteban" sort="Vzquez Fernndez, Esteban" uniqKey="Vzquez Fernndez E" first="Esteban" last="Vzquez-Fernndez">Esteban Vzquez-Fernndez</name><affiliation><mods:affiliation>Laboratorio Oficial de Metroloxa de Galicia (LOMG), Parque Tecnolxico de Galicia, San Cibrao das Vias, 32901, Ourense, Spain</mods:affiliation></affiliation><affiliation><mods:affiliation>E-mail:evazquez@lomg.net</mods:affiliation></affiliation></author><author><name sortKey="Dacal Nieto, Angel" sort="Dacal Nieto, Angel" uniqKey="Dacal Nieto A" first="Angel" last="Dacal-Nieto">Angel Dacal-Nieto</name><affiliation><mods:affiliation>Laboratorio Oficial de Metroloxa de Galicia (LOMG), Parque Tecnolxico de Galicia, San Cibrao das Vias, 32901, Ourense, Spain</mods:affiliation></affiliation></author><author><name sortKey="Gonzlez Jorge, Higinio" sort="Gonzlez Jorge, Higinio" uniqKey="Gonzlez Jorge H" first="Higinio" last="Gonzlez-Jorge">Higinio Gonzlez-Jorge</name><affiliation><mods:affiliation>Laboratorio Oficial de Metroloxa de Galicia (LOMG), Parque Tecnolxico de Galicia, San Cibrao das Vias, 32901, Ourense, Spain</mods:affiliation></affiliation></author><author><name sortKey="Martn, Fernando" sort="Martn, Fernando" uniqKey="Martn F" first="Fernando" last="Martn">Fernando Martn</name><affiliation><mods:affiliation>Departamento de Teora do Sinal e Comunicacins, Universidade de Vigo, Campus Universitario, 36310, Vigo, Spain</mods:affiliation></affiliation></author><author><name sortKey="Formella, Arno" sort="Formella, Arno" uniqKey="Formella A" first="Arno" last="Formella">Arno Formella</name><affiliation><mods:affiliation>Departamento de Informtica, Universidade de Vigo, As Lagoas s/n, 32004, Ourense, Spain</mods:affiliation></affiliation></author><author><name sortKey="Alvarez Valado, Victor" sort="Alvarez Valado, Victor" uniqKey="Alvarez Valado V" first="Victor" last="Alvarez-Valado">Victor Alvarez-Valado</name><affiliation><mods:affiliation>Laboratorio Oficial de Metroloxa de Galicia (LOMG), Parque Tecnolxico de Galicia, San Cibrao das Vias, 32901, Ourense, Spain</mods:affiliation></affiliation></author></titleStmt><publicationStmt><idno type="wicri:source">ISTEX</idno><idno type="RBID">ISTEX:EBF303D3EB21A4EED751C98AB5B05281A086F542</idno><date when="2009" year="2009">2009</date><idno type="doi">10.1088/0957-0233/20/6/065106</idno><idno type="url">https://api.istex.fr/document/EBF303D3EB21A4EED751C98AB5B05281A086F542/fulltext/pdf</idno><idno type="wicri:Area/Istex/Corpus">000503</idno></publicationStmt><sourceDesc><biblStruct><analytic><title level="a">A machine vision system for the calibration of digital thermometers</title><author><name sortKey="Vzquez Fernndez, Esteban" sort="Vzquez Fernndez, Esteban" uniqKey="Vzquez Fernndez E" first="Esteban" last="Vzquez-Fernndez">Esteban Vzquez-Fernndez</name><affiliation><mods:affiliation>Laboratorio Oficial de Metroloxa de Galicia (LOMG), Parque Tecnolxico de Galicia, San Cibrao das Vias, 32901, Ourense, Spain</mods:affiliation></affiliation><affiliation><mods:affiliation>E-mail:evazquez@lomg.net</mods:affiliation></affiliation></author><author><name sortKey="Dacal Nieto, Angel" sort="Dacal Nieto, Angel" uniqKey="Dacal Nieto A" first="Angel" last="Dacal-Nieto">Angel Dacal-Nieto</name><affiliation><mods:affiliation>Laboratorio Oficial de Metroloxa de Galicia (LOMG), Parque Tecnolxico de Galicia, San Cibrao das Vias, 32901, Ourense, Spain</mods:affiliation></affiliation></author><author><name sortKey="Gonzlez Jorge, Higinio" sort="Gonzlez Jorge, Higinio" uniqKey="Gonzlez Jorge H" first="Higinio" last="Gonzlez-Jorge">Higinio Gonzlez-Jorge</name><affiliation><mods:affiliation>Laboratorio Oficial de Metroloxa de Galicia (LOMG), Parque Tecnolxico de Galicia, San Cibrao das Vias, 32901, Ourense, Spain</mods:affiliation></affiliation></author><author><name sortKey="Martn, Fernando" sort="Martn, Fernando" uniqKey="Martn F" first="Fernando" last="Martn">Fernando Martn</name><affiliation><mods:affiliation>Departamento de Teora do Sinal e Comunicacins, Universidade de Vigo, Campus Universitario, 36310, Vigo, Spain</mods:affiliation></affiliation></author><author><name sortKey="Formella, Arno" sort="Formella, Arno" uniqKey="Formella A" first="Arno" last="Formella">Arno Formella</name><affiliation><mods:affiliation>Departamento de Informtica, Universidade de Vigo, As Lagoas s/n, 32004, Ourense, Spain</mods:affiliation></affiliation></author><author><name sortKey="Alvarez Valado, Victor" sort="Alvarez Valado, Victor" uniqKey="Alvarez Valado V" first="Victor" last="Alvarez-Valado">Victor Alvarez-Valado</name><affiliation><mods:affiliation>Laboratorio Oficial de Metroloxa de Galicia (LOMG), Parque Tecnolxico de Galicia, San Cibrao das Vias, 32901, Ourense, Spain</mods:affiliation></affiliation></author></analytic><monogr></monogr><series><title level="j">Measurement Science and Technology</title><title level="j" type="abbrev">Meas. Sci. Technol.</title><idno type="ISSN">0957-0233</idno><idno type="eISSN">1361-6501</idno><imprint><publisher>IOP Publishing</publisher><date type="published" when="2009">2009</date><biblScope unit="volume">20</biblScope><biblScope unit="issue">6</biblScope><biblScope unit="page" from="1">1</biblScope><biblScope unit="page" to="7">7</biblScope><biblScope unit="production">Printed in the UK</biblScope></imprint><idno type="ISSN">0957-0233</idno></series><idno type="istex">EBF303D3EB21A4EED751C98AB5B05281A086F542</idno><idno type="DOI">10.1088/0957-0233/20/6/065106</idno><idno type="PII">S0957-0233(09)05869-X</idno><idno type="articleID">305869</idno><idno type="articleNumber">065106</idno></biblStruct></sourceDesc><seriesStmt><idno type="ISSN">0957-0233</idno></seriesStmt></fileDesc><profileDesc><textClass></textClass><langUsage><language ident="en">en</language></langUsage></profileDesc></teiHeader><front><div type="abstract">Automation is a key point in many industrial tasks such as calibration and metrology. In this context, machine vision has shown to be a useful tool for automation support, especially when there is no other option available. A system for the calibration of portable measurement devices has been developed. The system uses machine vision to obtain the numerical values shown by displays. A new approach based on human perception of digits, which works in parallel with other more classical classifiers, has been created. The results show the benefits of the system in terms of its usability and robustness, obtaining a success rate higher than 99 in display recognition. The system saves time and effort, and offers the possibility of scheduling calibration tasks without excessive attention by the laboratory technicians.</div></front></TEI><istex><corpusName>iop</corpusName><author><json:item><name>Esteban Vzquez-Fernndez</name><affiliations><json:string>Laboratorio Oficial de Metroloxa de Galicia (LOMG), Parque Tecnolxico de Galicia, San Cibrao das Vias, 32901, Ourense, Spain</json:string><json:string>E-mail:evazquez@lomg.net</json:string></affiliations></json:item><json:item><name>Angel Dacal-Nieto</name><affiliations><json:string>Laboratorio Oficial de Metroloxa de Galicia (LOMG), Parque Tecnolxico de Galicia, San Cibrao das Vias, 32901, Ourense, Spain</json:string></affiliations></json:item><json:item><name>Higinio Gonzlez-Jorge</name><affiliations><json:string>Laboratorio Oficial de Metroloxa de Galicia (LOMG), Parque Tecnolxico de Galicia, San Cibrao das Vias, 32901, Ourense, Spain</json:string></affiliations></json:item><json:item><name>Fernando Martn</name><affiliations><json:string>Departamento de Teora do Sinal e Comunicacins, Universidade de Vigo, Campus Universitario, 36310, Vigo, Spain</json:string></affiliations></json:item><json:item><name>Arno Formella</name><affiliations><json:string>Departamento de Informtica, Universidade de Vigo, As Lagoas s/n, 32004, Ourense, Spain</json:string></affiliations></json:item><json:item><name>Victor Alvarez-Valado</name><affiliations><json:string>Laboratorio Oficial de Metroloxa de Galicia (LOMG), Parque Tecnolxico de Galicia, San Cibrao das Vias, 32901, Ourense, Spain</json:string></affiliations></json:item></author><subject><json:item><lang><json:string>eng</json:string></lang><value>metrology</value></json:item><json:item><lang><json:string>eng</json:string></lang><value>machine vision</value></json:item><json:item><lang><json:string>eng</json:string></lang><value>calibration</value></json:item><json:item><lang><json:string>eng</json:string></lang><value>thermometer</value></json:item><json:item><lang><json:string>eng</json:string></lang><value>automation</value></json:item></subject><language><json:string>eng</json:string></language><abstract>Automation is a key point in many industrial tasks such as calibration and metrology. In this context, machine vision has shown to be a useful tool for automation support, especially when there is no other option available. A system for the calibration of portable measurement devices has been developed. The system uses machine vision to obtain the numerical values shown by displays. A new approach based on human perception of digits, which works in parallel with other more classical classifiers, has been created. The results show the benefits of the system in terms of its usability and robustness, obtaining a success rate higher than 99 in display recognition. The system saves time and effort, and offers the possibility of scheduling calibration tasks without excessive attention by the laboratory technicians.</abstract><qualityIndicators><score>5.04</score><pdfVersion>1.4</pdfVersion><pdfPageSize>595 x 842 pts (A4)</pdfPageSize><refBibsNative>true</refBibsNative><keywordCount>5</keywordCount><abstractCharCount>820</abstractCharCount><pdfWordCount>2992</pdfWordCount><pdfCharCount>18294</pdfCharCount><pdfPageCount>7</pdfPageCount><abstractWordCount>129</abstractWordCount></qualityIndicators><title>A machine vision system for the calibration of digital thermometers</title><genre.original><json:string>paper</json:string></genre.original><pii><json:string>S0957-0233(09)05869-X</json:string></pii><genre><json:string>research-article</json:string></genre><host><volume>20</volume><publisherId><json:string>MST</json:string></publisherId><pages><total>7</total><last>7</last><first>1</first></pages><issn><json:string>0957-0233</json:string></issn><issue>6</issue><genre><json:string>journal</json:string></genre><language><json:string>unknown</json:string></language><eissn><json:string>1361-6501</json:string></eissn><title>Measurement Science and Technology</title></host><publicationDate>2009</publicationDate><copyrightDate>2009</copyrightDate><doi><json:string>10.1088/0957-0233/20/6/065106</json:string></doi><id>EBF303D3EB21A4EED751C98AB5B05281A086F542</id><fulltext><json:item><original>true</original><mimetype>application/pdf</mimetype><extension>pdf</extension><uri>https://api.istex.fr/document/EBF303D3EB21A4EED751C98AB5B05281A086F542/fulltext/pdf</uri></json:item><json:item><original>false</original><mimetype>application/zip</mimetype><extension>zip</extension><uri>https://api.istex.fr/document/EBF303D3EB21A4EED751C98AB5B05281A086F542/fulltext/zip</uri></json:item><istex:fulltextTEI uri="https://api.istex.fr/document/EBF303D3EB21A4EED751C98AB5B05281A086F542/fulltext/tei"><teiHeader><fileDesc><titleStmt><title level="a">A machine vision system for the calibration of digital thermometers</title></titleStmt><publicationStmt><authority>ISTEX</authority><publisher>IOP Publishing</publisher><availability><p>IOP</p></availability><date>2009</date></publicationStmt><sourceDesc><biblStruct type="inbook"><analytic><title level="a">A machine vision system for the calibration of digital thermometers</title><author><persName><forename type="first">Esteban</forename><surname>Vzquez-Fernndez</surname></persName><affiliation>Laboratorio Oficial de Metroloxa de Galicia (LOMG), Parque Tecnolxico de Galicia, San Cibrao das Vias, 32901, Ourense, Spain</affiliation><affiliation>E-mail:evazquez@lomg.net</affiliation></author><author><persName><forename type="first">Angel</forename><surname>Dacal-Nieto</surname></persName><affiliation>Laboratorio Oficial de Metroloxa de Galicia (LOMG), Parque Tecnolxico de Galicia, San Cibrao das Vias, 32901, Ourense, Spain</affiliation></author><author><persName><forename type="first">Higinio</forename><surname>Gonzlez-Jorge</surname></persName><affiliation>Laboratorio Oficial de Metroloxa de Galicia (LOMG), Parque Tecnolxico de Galicia, San Cibrao das Vias, 32901, Ourense, Spain</affiliation></author><author><persName><forename type="first">Fernando</forename><surname>Martn</surname></persName><affiliation>Departamento de Teora do Sinal e Comunicacins, Universidade de Vigo, Campus Universitario, 36310, Vigo, Spain</affiliation></author><author><persName><forename type="first">Arno</forename><surname>Formella</surname></persName><affiliation>Departamento de Informtica, Universidade de Vigo, As Lagoas s/n, 32004, Ourense, Spain</affiliation></author><author><persName><forename type="first">Victor</forename><surname>Alvarez-Valado</surname></persName><affiliation>Laboratorio Oficial de Metroloxa de Galicia (LOMG), Parque Tecnolxico de Galicia, San Cibrao das Vias, 32901, Ourense, Spain</affiliation></author></analytic><monogr><title level="j">Measurement Science and Technology</title><title level="j" type="abbrev">Meas. Sci. Technol.</title><idno type="pISSN">0957-0233</idno><idno type="eISSN">1361-6501</idno><imprint><publisher>IOP Publishing</publisher><date type="published" when="2009"></date><biblScope unit="volume">20</biblScope><biblScope unit="issue">6</biblScope><biblScope unit="page" from="1">1</biblScope><biblScope unit="page" to="7">7</biblScope><biblScope unit="production">Printed in the UK</biblScope></imprint></monogr><idno type="istex">EBF303D3EB21A4EED751C98AB5B05281A086F542</idno><idno type="DOI">10.1088/0957-0233/20/6/065106</idno><idno type="PII">S0957-0233(09)05869-X</idno><idno type="articleID">305869</idno><idno type="articleNumber">065106</idno></biblStruct></sourceDesc></fileDesc><profileDesc><creation><date>2009</date></creation><langUsage><language ident="en">en</language></langUsage><abstract><p>Automation is a key point in many industrial tasks such as calibration and metrology. In this context, machine vision has shown to be a useful tool for automation support, especially when there is no other option available. A system for the calibration of portable measurement devices has been developed. The system uses machine vision to obtain the numerical values shown by displays. A new approach based on human perception of digits, which works in parallel with other more classical classifiers, has been created. The results show the benefits of the system in terms of its usability and robustness, obtaining a success rate higher than 99 in display recognition. The system saves time and effort, and offers the possibility of scheduling calibration tasks without excessive attention by the laboratory technicians.</p></abstract><textClass><keywords scheme="keyword"><list><head>keywords</head><item><term>metrology</term></item><item><term>machine vision</term></item><item><term>calibration</term></item><item><term>thermometer</term></item><item><term>automation</term></item></list></keywords></textClass><textClass><classCode scheme="">06.20.-f</classCode><classCode scheme="">06.20.fb</classCode><classCode scheme="">06.60.Mr</classCode><classCode scheme="">07.20.Dt</classCode></textClass></profileDesc><revisionDesc><change when="2009">Published</change></revisionDesc></teiHeader></istex:fulltextTEI><json:item><original>false</original><mimetype>text/plain</mimetype><extension>txt</extension><uri>https://api.istex.fr/document/EBF303D3EB21A4EED751C98AB5B05281A086F542/fulltext/txt</uri></json:item></fulltext><metadata><istex:metadataXml wicri:clean="corpus iop not found" wicri:toSee="no header"><istex:xmlDeclaration>version="1.0" encoding="ISO-8859-1"</istex:xmlDeclaration><istex:docType SYSTEM="http://ej.iop.org/dtd/iopv1_5_2.dtd" name="istex:docType"></istex:docType><istex:document><article artid="mst305869"><article-metadata><jnl-data jnlid="mst"><jnl-fullname>Measurement Science and Technology</jnl-fullname><jnl-abbreviation>Meas. Sci. Technol.</jnl-abbreviation><jnl-shortname>MST</jnl-shortname><jnl-issn>0957-0233</jnl-issn><jnl-coden>MSTCEP</jnl-coden><jnl-imprint>IOP Publishing</jnl-imprint><jnl-web-address>stacks.iop.org/MST</jnl-web-address></jnl-data><volume-data><year-publication>2009</year-publication><volume-number>20</volume-number></volume-data><issue-data><issue-number>6</issue-number><coverdate>June 2009</coverdate></issue-data><article-data><article-type type="paper" sort="regular"></article-type><type-number type="paper" numbering="article" artnum="065106">065106</type-number><article-number>305869</article-number><first-page>1</first-page><last-page>7</last-page><length>7</length><pii>S0957-0233(09)05869-X</pii><doi>10.1088/0957-0233/20/6/065106</doi><copyright>2009 IOP Publishing Ltd</copyright><ccc>0957-0233/09/065106+07$30.00</ccc><printed>Printed in the UK</printed><features colour="global" mmedia="no" suppdata="no"></features></article-data></article-metadata><header><title-group><title>A machine vision system for the calibration of digital thermometers</title><short-title>A machine vision system for the calibration of digital thermometers</short-title><ej-title>A machine vision system for the calibration of digital thermometers</ej-title></title-group><author-group><author address="mst305869ad1" email="mst305869ea1"><first-names>Esteban</first-names><second-name>Vázquez-Fernández</second-name></author><author address="mst305869ad1"><first-names>Angel</first-names><second-name>Dacal-Nieto</second-name></author><author address="mst305869ad1"><first-names>Higinio</first-names><second-name>González-Jorge</second-name></author><author address="mst305869ad2"><first-names>Fernando</first-names><second-name>Martín</second-name></author><author address="mst305869ad3"><first-names>Arno</first-names><second-name>Formella</second-name></author><author address="mst305869ad1"><first-names>Victor</first-names><second-name>Alvarez-Valado</second-name></author></author-group><address-group><address id="mst305869ad1"><orgname>Laboratorio Oficial de Metroloxía de Galicia (LOMG), Parque Tecnolóxico de Galicia</orgname>, San Cibrao das Viñas, 32901, Ourense, <country>Spain</country></address><address id="mst305869ad2"><orgname>Departamento de Teoría do Sinal e Comunicacións, Universidade de Vigo</orgname>, Campus Universitario, 36310, Vigo, <country>Spain</country></address><address id="mst305869ad3"><orgname>Departamento de Informática, Universidade de Vigo</orgname>, As Lagoas s/n, 32004, Ourense, <country>Spain</country></address><e-address id="mst305869ea1"><email mailto="evazquez@lomg.net">evazquez@lomg.net</email></e-address></address-group><history received="14 January 2009" finalform="26 March 2009" online="18 May 2009"></history><abstract-group><abstract><heading>Abstract</heading><p indent="no">Automation is a key point in many industrial tasks such as calibration and metrology. In this context, machine vision has shown to be a useful tool for automation support, especially when there is no other option available. A system for the calibration of portable measurement devices has been developed. The system uses machine vision to obtain the numerical values shown by displays. A new approach based on human perception of digits, which works in parallel with other more classical classifiers, has been created. The results show the benefits of the system in terms of its usability and robustness, obtaining a success rate higher than 99% in display recognition. The system saves time and effort, and offers the possibility of scheduling calibration tasks without excessive attention by the laboratory technicians.</p></abstract></abstract-group><classifications><class-codes scheme="pacs"><code>06.20.-f</code><code>06.20.fb</code><code>06.60.Mr</code><code>07.20.Dt</code></class-codes><keywords print="yes"><keyword>metrology</keyword><keyword>machine vision</keyword><keyword>calibration</keyword><keyword>thermometer</keyword><keyword>automation</keyword></keywords></classifications></header><body numbering="bysection"><sec-level1 id="mst305869s1" label="1"><heading>Introduction</heading><p indent="no">Automation in calibration processes offers a great number of advantages as in speed, in accuracy or in decreasing the error rate [<cite linkend="mst305869bib01">1</cite>]. Optimization of these processes is a very important task and, in this field, machine vision systems play a fundamental role [<cite linkend="mst305869bib02" range="mst305869bib02,mst305869bib03,mst305869bib04">2–4</cite>]. In the daily work of a metrology laboratory many instruments need to be periodically inspected at the times required by a calibration protocol. Writing down these measurements is often done manually by the lab technicians. Such a procedure requires full-time work and leads to undesirable delays in calibration in many cases. To make this routine work easier, a machine vision system was developed.</p><p>Optical character recognition (OCR) is a pattern recognition problem which has been studied for over 60 years. Multiple classifiers and classification schemes, feature extraction methods, etc have been developed for countless applications in many fields. One can find a lot of literature on these issues from different perspectives [<cite linkend="mst305869bib05">5</cite>]. Commercial OCR packages are also available, but most are mainly oriented to document reading and interpretation (not real world scenes). In these applications there are requirements in the capture, illumination, positioning, etc which we cannot achieve, due to the restrictive environment of the calibration process. We have tried GOCR, an OCR program developed under GNU license (<webref url="http://www.jocr.sourceforge.net">www.jocr.sourceforge.net</webref>), without success (10% recognition rate). There are also other tools available such as machine vision interfaces or expert systems similar to automatic car-plate recognition systems, which are capable of working under difficult capture conditions. However, these kinds of OCR are designed to work with a previous known format for the digit representation.</p><p>There is also some related work dealing with the instrument-reading problem [<cite linkend="mst305869bib06">6</cite>], but it focuses merely on a previous known single font display (usually seven-segment). However, common instruments have multiple different fonts and may have defects such as scratches or bubbles due to their use in industrial environments, which increases the variance of characters and makes a correct reading more difficult. Our goal is to develop a sufficiently reliable system able to read almost any display without any non-obvious previous assumption about its font type.</p><p>The work has been developed for supporting and improving the calibration processes of the Temperature and Humidity Department of the ‘Laboratorio Oficial de Metroloxía de Galicia, LOMG’ (Spain). In this application, the instruments under calibration are mainly portable thermometers and hygrometers (see figure <figref linkend="mst305869fig01">1</figref>).
<figure id="mst305869fig01"><graphic><graphic-file version="print" format="EPS" filename="images/mst305869fig01.eps" width="18pc"></graphic-file><graphic-file version="ej" format="JPEG" filename="images/mst305869fig01.jpg"></graphic-file></graphic><caption id="mst305869fc01" label="Figure 1"><p indent="no">Some samples of instruments.</p></caption></figure></p><p>The paper is organized as follows. Section <secref linkend="mst305869s2">2</secref> provides a system overview. Then, section <secref linkend="mst305869s3">3</secref> shows the image processing algorithms. Finally, some results and discussion are presented.</p></sec-level1><sec-level1 id="mst305869s2" label="2"><heading>System overview</heading><p indent="no">A temperature calibration process establishes the comparison between the measurements obtained by a pattern thermometer and a measurand thermometer under controlled environmental conditions (provided by devices such as baths, climatic chambers, etc). This procedure ensures the traceability of the measurement in the standard definition of Kelvin.</p><p>Calibrations in the Temperature and Humidity Department had been partially automated using the commercial software provided by the pattern devices and the baths for their management (through RS-232 and GPIB ports). Nevertheless, to achieve a better level of automation, we should communicate not only with the pattern devices and baths but also with the measurands. The problem arises that most of the instruments do not have a communication port and there is no standard protocol for this purpose in the remaining ones. Under these conditions, machine vision becomes a key factor for reading and automation.</p><p>In most cases measurands have an LCD display where the measurements are shown. The idea is to automatically obtain images of these instruments, extract the numerical values that are shown on the displays and communicate with pattern devices and baths when the protocol requires it. A schematic view is shown in figure <figref linkend="mst305869fig02">2</figref>.
<figure id="mst305869fig02"><graphic><graphic-file version="print" format="EPS" filename="images/mst305869fig02.eps" width="20.5pc"></graphic-file><graphic-file version="ej" format="JPEG" filename="images/mst305869fig02.jpg"></graphic-file></graphic><caption id="mst305869fc02" label="Figure 2"><p indent="no">Scheme of a thermometer calibration system.</p></caption></figure></p><p>We achieved the best results with a C-Cam BCi4 CMOS (C-Cam Tech.: <webref url="http://www.c-cam.be">www.c-cam.be</webref>) camera at 1280 × 1024 resolution. A 16 mm lens has shown to be enough to cover the field of view of a typical display at a close distance (20–25 cm). However, a standard 640 × 480 webcam can be used, accepting a decrease in the correct recognition rate. A comparison between the two camera systems is shown in section <secref linkend="mst305869s4">4</secref>.</p><p>Another important improvement of the system, compared to the manual procedure, is the possibility to program a sequence of steps for different temperature ranges in the calibration baths, so that the system can be scheduled to work for 24 h a day without supervision. A manual calibration process of thermometers is long and tedious; with this improvement, the efficiency in the laboratory has increased significantly.</p><p>In addition, an assembly (figure <figref linkend="mst305869fig03">3</figref>) has been designed to fix the camera–thermometer system, so that it is possible to move the measurand from one bath to another with no additional re-configurations.
<figure id="mst305869fig03"><graphic><graphic-file version="print" format="EPS" filename="images/mst305869fig03.eps" width="20.5pc"></graphic-file><graphic-file version="ej" format="JPEG" filename="images/mst305869fig03.jpg"></graphic-file></graphic><caption id="mst305869fc03" label="Figure 3"><p indent="no">Mechanical mounting.</p></caption></figure></p><p>Before starting a calibration process, it is necessary to provide the system with some information relative to the devices to be calibrated, protocols to be followed, etc. The area where the display is located on an image must be found only once, because the setup camera device is fixed. Currently, we still rely on the user to select the region where the display is located. The graphical user interface (GUI) developed to allow these functions is shown in figure <figref linkend="mst305869fig04">4</figref>.
<figure id="mst305869fig04" width="page"><graphic><graphic-file version="print" format="EPS" filename="images/mst305869fig04.eps" width="26pc"></graphic-file><graphic-file version="ej" format="JPEG" filename="images/mst305869fig04.jpg"></graphic-file></graphic><caption id="mst305869fc04" label="Figure 4"><p indent="no">System GUI (region of interest is marked with a rectangle).</p></caption></figure></p><p>The user can define more than one region of interest (ROI), a useful feature in the case of instruments which have several indications in their displays, or in the case that several instruments can be calibrated in one experiment. The latter happens, for instance, with thermohygrometers, where two values, namely the temperature and the relative humidity, are read. Other parameters of the camera such as the capture period are also selected in this initial phase.</p></sec-level1><sec-level1 id="mst305869s3" label="3"><heading>Image processing</heading><p indent="no">The ROI extracted by the user is analyzed with standard image processing techniques as described in some detail next. We have enhanced our system with a hybrid recognizer scheme, which combines the use of a classical distance classifier with a new recognizer based on simple human feature perception.</p><p>The recognition process starts with the binarization of the image. To avoid problems in images with important illumination gradients, we have designed a multistep procedure. To separate background and foreground in the ROI, a method based on searching the histogram peaks and locating thresholds on the minima between them is used [<cite linkend="mst305869bib07">7</cite>]. This method is only effective in the case of bimodal histograms, so if there are more than two prominent peaks, the algorithm switches to an alternative method. First, the threshold quality is estimated with the measurement of the area in the threshold boundary (figure <figref linkend="mst305869fig05">5</figref>). This value is an indicator of the separability of the peaks in a bimodal histogram. If the area exceeds 2.5% of the entire histogram, we switch to a threshold algorithm based on interpolation [<cite linkend="mst305869bib08">8</cite>] with a 4 × 3 sub-image grid which uses the Otsu threshold on each piece which is implemented via an approximate iterative version [<cite linkend="mst305869bib09">9</cite>, <cite linkend="mst305869bib10">10</cite>]. The overall procedure is summarized in figure <figref linkend="mst305869fig06">6</figref>.
<figure id="mst305869fig05"><graphic><graphic-file version="print" format="EPS" filename="images/mst305869fig05.eps" width="20.5pc"></graphic-file><graphic-file version="ej" format="JPEG" filename="images/mst305869fig05.jpg"></graphic-file></graphic><caption id="mst305869fc05" label="Figure 5"><p indent="no">Peak detection method: the area in the threshold boundary measures the quality.</p></caption></figure><figure id="mst305869fig06"><graphic><graphic-file version="print" format="EPS" filename="images/mst305869fig06.eps" width="12pc"></graphic-file><graphic-file version="ej" format="JPEG" filename="images/mst305869fig06.jpg"></graphic-file></graphic><caption id="mst305869fc06" label="Figure 6"><p indent="no">Binarization process.</p></caption></figure></p><p>Then, other typical preprocessing techniques, including filters to reduce noise or skew angle correction (figure <figref linkend="mst305869fig07">7</figref>), are applied to enhance the image before the segmentation process starts. Character row extraction is also applied taking advantage of big values on horizontal and vertical image projections to detect the edges.
<figure id="mst305869fig07"><graphic><graphic-file version="print" format="EPS" filename="images/mst305869fig07.eps" width="18pc"></graphic-file><graphic-file version="ej" format="JPEG" filename="images/mst305869fig07.jpg"></graphic-file></graphic><caption id="mst305869fc07" label="Figure 7"><p indent="no">Angle correction: (<italic>a</italic>) method overview; (<italic>b</italic>) skewed image example; (<italic>c</italic>) extracted contour (vertical transition points); (<italic>d</italic>) cleaned contour after erasing points far from the mean and isolated ones; (<italic>e</italic>) corrected image.</p></caption></figure></p><p>A commonly used technique for character segmentation in a preprocessed row is based on horizontal projections [<cite linkend="mst305869bib11">11</cite>]. Keeping in mind that we have to deal with instruments used in industrial environments, we apply an enhanced projection technique which is found to be more effective in noisy images. The aim of the method is to compute the position <italic>i</italic> in the projection vector as the dot product between the (<italic>i</italic> − 1)st and (<italic>i</italic> + 1)st column. The benefits of this technique in segmentation are shown in figure <figref linkend="mst305869fig08">8</figref>, obtaining deeper projection minima.
<figure id="mst305869fig08" width="page"><graphic><graphic-file version="print" format="EPS" filename="images/mst305869fig08.eps" width="31pc"></graphic-file><graphic-file version="ej" format="JPEG" filename="images/mst305869fig08.jpg"></graphic-file></graphic><caption id="mst305869fc08" label="Figure 8"><p indent="no">Comparison of (<italic>a</italic>) standard projection and (<italic>b</italic>) enhanced projection.</p></caption></figure></p><p>The segmentation process is complemented with a verification technique that checks, e.g., the aspect ratio or number of peaks and valleys of segmented character projections to detect linked characters (figure <figref linkend="mst305869fig09">9</figref>).
<figure id="mst305869fig09"><graphic><graphic-file version="print" format="EPS" filename="images/mst305869fig09.eps" width="20.5pc"></graphic-file><graphic-file version="ej" format="JPEG" filename="images/mst305869fig09.jpg"></graphic-file></graphic><caption id="mst305869fc09" label="Figure 9"><p indent="no">Row projection: a correctly segmented character has a maximum of two local valleys.</p></caption></figure></p><p>A redundant search of the decimal point is also applied to deal with problems like skewed characters that confuse the original projection. In the case of not finding the point in the first run, we start a search using a projection of only the lower part of the characters being already segmented (figure <figref linkend="mst305869fig10">10</figref>).
<figure id="mst305869fig10"><graphic><graphic-file version="print" format="EPS" filename="images/mst305869fig10.eps" width="20.5pc"></graphic-file><graphic-file version="ej" format="JPEG" filename="images/mst305869fig10.jpg"></graphic-file></graphic><caption id="mst305869fc10" label="Figure 10"><p indent="no">Above: skewed characters that make the point detection difficult. Below left: lower part of ‘2’. Below right: projection of fragment.</p></caption></figure></p><p>For digit recognition, we take advantage of two different classifiers. The first one is a standard structured classifier based on feature extraction and distance calculation. The selected features are character projections and directional components based on Kirsch Gradients [<cite linkend="mst305869bib12">12</cite>]. For the classifier, a 1-nearest-neighbor algorithm is used. We have also tried various other classifiers (such as probabilistic neural networks, Gaussian classifiers and k-NN) but the best results were achieved for the 1-NN. The reason seems to be our multi-font situation, where the intra-class variance is greater than the inter-class variance for many cases.</p><p>As patterns for the 1-NN classifier we have chosen perfect ones (obtained from different computer font types such as Arial, Times, Digital, etc). We tried to use patterns from segmented input digits, but the 1-NN got better results for the artificial, perfect ones.</p><p>The second classifier is an original design based on human feature perception [<cite linkend="mst305869bib13">13</cite>]. The objective is to inspect geometric features such as the presence of holes, lines or openings in variable positions around seven non-fixed outstanding regions in the character geometry (figure <figref linkend="mst305869fig11">11</figref>). These regions are compared to the known expectations for every character from 0 to 9. We obtain a ten-component binary vector (0–9) which has the value 0 (zero) or 1 (one) for each class, depending on its compatibility with the observed characteristics.
<figure id="mst305869fig11"><graphic><graphic-file version="print" format="EPS" filename="images/mst305869fig11.eps" width="10pc"></graphic-file><graphic-file version="ej" format="JPEG" filename="images/mst305869fig11.jpg"></graphic-file></graphic><caption id="mst305869fc11" label="Figure 11"><p indent="no">Areas of interest.</p></caption></figure></p><p>The aim of this method is not to serve as a global classifier able to work by itself, rather than to cooperate with another kind of classifier. The key is to differentiate the cases which are problematic for the first classifier. For this purpose, the feature perception method and the 1-NN classifier are combined in the way shown in figure <figref linkend="mst305869fig12">12</figref>. However, a similar fusion scheme can be used with any other kind of classifier.
<figure id="mst305869fig12"><graphic><graphic-file version="print" format="EPS" filename="images/mst305869fig12.eps" width="17.5pc"></graphic-file><graphic-file version="ej" format="JPEG" filename="images/mst305869fig12.jpg"></graphic-file></graphic><caption id="mst305869fc12" label="Figure 12"><p indent="no">Classifier fusion.</p></caption></figure></p><p>The summary of the classification procedure is as follows.
<itemized-list id="mst305869il1"><list-item id="mst305869il1.1" marker="•"><p indent="no">On the one hand, feature extraction and norm-1 distance to every pattern are computed to obtain a distance vector.</p></list-item><list-item id="mst305869il1.2" marker="•"><p indent="no">On the other hand, the feature perception method, which results in a recognized character (or several possible characters), gives us a binary vector where its positions correspond to matching classes coded as ones and others coded as zeros.</p></list-item><list-item id="mst305869il1.3" marker="•"><p indent="no">Distances corresponding to the classes recognized by the feature perception method are reduced by 20% (empirically) in the distance vector.</p></list-item><list-item id="mst305869il1.4" marker="•"><p indent="no">Finally, the 1-NN is applied.</p></list-item></itemized-list></p></sec-level1><sec-level1 id="mst305869s4" label="4"><heading>Results and discussion</heading><p indent="no">To test the behavior of the system under real operation conditions, we have used 448 images obtained from 16 sequences of different instruments (figure <figref linkend="mst305869fig13">13</figref>). Note that we only take one sequence for every display. The system obtained the correct values 445 times, i.e. 99.33% recognition rate (measured on display images, not on individual digits). These results are achieved using the C-Cam BCi4 1280 × 1024 camera.
<figure id="mst305869fig13"><graphic><graphic-file version="print" format="EPS" filename="images/mst305869fig13.eps" width="17.5pc"></graphic-file><graphic-file version="ej" format="JPEG" filename="images/mst305869fig13.jpg"></graphic-file></graphic><caption id="mst305869fc13" label="Figure 13"><p indent="no">Different fonts of the displays: (<italic>a</italic>) seven-segment skewed/not skewed, (<italic>b</italic>) effects of bubbles and scratched displays, (<italic>c</italic>) and graphic displays.</p></caption></figure></p><p>A test using a standard 640 × 480 pixel webcam (Labtec Webcam 1200, <webref url="http://www.labtec.com">www.labtec.com</webref>) was also made. The achieved correct recognition rate is slightly lower (table <tabref linkend="mst305869tab01">1</tabref> and figure <figref linkend="mst305869fig14">14</figref>) in our test setup. In a real operation scenario, however, a better recognition rate can be expected due to the fact that various images can be shot in the stable region of the measurements which is the case in a typical calibration process where the measurements are made under controlled conditions and when the instruments have achieved a stability point. At this time, several redundant measurements can be made, providing enough information to discard possible incorrect interpretations.
<figure id="mst305869fig14"><graphic><graphic-file version="print" format="EPS" filename="images/mst305869fig14.eps" width="20.5pc"></graphic-file><graphic-file version="ej" format="JPEG" filename="images/mst305869fig14.jpg"></graphic-file></graphic><caption id="mst305869fc14" label="Figure 14"><p indent="no">Camera performance comparison.</p></caption></figure><table id="mst305869tab01" frame="topbot"><caption id="mst305869tc01" label="Table 1"><p indent="no">Camera performance.</p></caption><tgroup cols="5"><colspec colnum="1" colname="col1" align="left"></colspec><colspec colnum="2" colname="col2" align="left"></colspec><colspec colnum="3" colname="col3" align="left"></colspec><colspec colnum="4" colname="col4" align="left"></colspec><colspec colnum="5" colname="col5" align="left"></colspec><thead><row><entry></entry><entry>Total</entry><entry>Correct</entry><entry>Incorrect</entry><entry>Correct (%)</entry></row></thead><tbody><row><entry>C-Cam BCi4</entry><entry>448</entry><entry>445</entry><entry>3</entry><entry>99.33</entry></row><row><entry>Webcam</entry><entry>194</entry><entry>185</entry><entry>9</entry><entry>95.36</entry></row></tbody></tgroup></table></p><p>Errors are produced mainly in instruments with damaged displays (e.g. exhibiting scratches). Another cause of error is the capture of unstable displays where the digits are changing from one value to another (figure <figref linkend="mst305869fig15">15</figref>).
<figure id="mst305869fig15"><graphic><graphic-file version="print" format="EPS" filename="images/mst305869fig15.eps" width="20.5pc"></graphic-file><graphic-file version="ej" format="JPEG" filename="images/mst305869fig15.jpg"></graphic-file></graphic><caption id="mst305869fc15" label="Figure 15"><p indent="no">Problematic displays: (<italic>a</italic>) special display; note F above the decimals and (<italic>b</italic>) display in transition.</p></caption></figure></p><p>As explained before, the objective of our system is to be able to read any kind of display without previous knowledge about its font type. In practice, there appear additional problems like unexpected types of displays, not disabled automatic shutdown function, etc. Under such conditions, there will always be a certain percentage of instruments that cannot be read by the system.</p><p>The impossibility of automatic reading is noticed by the laboratory technician in the initial phase of the calibration, so it will not produce errors in interpretation. Problematic instruments can be scheduled to be calibrated under manual supervision while the other ones can be handled automatically. It is important to take this fact into account since it is one key in the robustness of the system.</p><p>We have checked the instruments scheduled for calibration in the Temperature Department over a period of 4 weeks. We have contrasted the results of these tests with the historic data of thermometers and hygrometers of the years 2007 and 2008. Using this information, we have an estimation of the instruments that allow or do not allow automatic reading by the machine vision system. The instruments which do not allow automation are approximately 13% of the total (table <tabref linkend="mst305869tab02">2</tabref> and figure <figref linkend="mst305869fig16">16</figref>). Note that auto shutdown is a problem beyond our system and should not be considered as an error.
<figure id="mst305869fig16" width="page"><graphic><graphic-file version="print" format="EPS" filename="images/mst305869fig16.eps" width="21pc"></graphic-file><graphic-file version="ej" format="JPEG" filename="images/mst305869fig16.jpg"></graphic-file></graphic><caption id="mst305869fc16" label="Figure 16"><p indent="no">Estimation of the system usability.</p></caption></figure><table id="mst305869tab02" frame="topbot"><caption id="mst305869tc02" label="Table 2"><p indent="no">System usability.</p></caption><tgroup cols="3"><colspec colnum="1" colname="col1" align="left"></colspec><colspec colnum="2" colname="col2" align="left"></colspec><colspec colnum="3" colname="col3" align="left"></colspec><thead><row><entry></entry><entry>Meters</entry><entry>(%)</entry></row></thead><tbody><row><entry>Readable instruments</entry><entry>428</entry><entry> 86.99</entry></row><row><entry>Incompatible format</entry><entry> 10</entry><entry> 2.03</entry></row><row><entry>Auto shutdown</entry><entry> 31</entry><entry> 6.30</entry></row><row><entry>Difficult displays</entry><entry> 23</entry><entry> 4.67</entry></row><row><entry>Total</entry><entry>492</entry><entry>100</entry></row></tbody></tgroup></table></p><p>Once tested for the thermometer calibration application, the system seems versatile enough to be adapted to read any kind of instruments exhibiting a numerical display. Consequently, other potential applications, like the reading of Doppler speed meters in a running car for calibration purposes, are being implemented at this time.</p><p>There are also some future lines we are working on to improve the system.
<itemized-list id="mst305869il2"><list-item id="mst305869il2.1" marker="•"><p indent="no">We could use sequences of images to solve the problem of displays in transition.</p></list-item><list-item id="mst305869il2.2" marker="•"><p indent="no">We are also studying the effects (in global recognition rate) of adding new fonts to treat incompatible formats.</p></list-item><list-item id="mst305869il2.3" marker="•"><p indent="no">The use of different preprocessing techniques, such as mathematical morphology or colorimetry, could be helpful to deal with difficult displays (e.g. damaged digits).</p></list-item><list-item id="mst305869il2.4" marker="•"><p indent="no">We are working on enhancing the detection of the decimal point in the case of different locations (centered, upper position, etc) or different symbols (point, dash, etc).</p></list-item></itemized-list></p></sec-level1><sec-level1 id="mst305869s5" label="5"><heading>Conclusions</heading><p indent="no">We have designed and implemented a working system which is able to read almost any display of digital instrumentation. There are many potential applications where this system can be a helpful tool and provide benefits such as saving time and effort and avoiding human errors. The system has shifted the working plans of the technicians toward less routine work and more qualified activities. It allows automatic 24 h processes to be scheduled.</p><p>Related to the image processing and pattern recognition work, the analysis of human behavior can be very helpful in improving processes where machine vision is applied. The combination of a standard structured classifier and a new recognizer based on human feature perception achieves a much stronger system, which is capable of working with a large intra-class variance due to the presence of multiple fonts.</p></sec-level1><acknowledgment><heading>Acknowledgments</heading><p indent="no">The authors would like to thank F J Yebra group at Temperature and Humidity Department (LOMG) for their helpful collaboration and ‘Xunta de Galicia’ for the financial support (project code 07TIC008CT and human resources program Lucas Labrada). The project was partially funded by the University of Vigo, project number 2008-INOU-5.</p></acknowledgment></body><back><references><heading>References</heading><reference-list type="numeric"><journal-ref id="mst305869bib01"><authors><au><second-name>Gurevitch</second-name><first-names>D</first-names></au></authors><year>2004</year><art-title>Economic justification of laboratory automation</art-title><jnl-title>J. Assoc. Lab. Autom.</jnl-title><volume>9</volume><pages>33–43</pages><crossref><cr_doi>http://dx.doi.org/10.1016/S1535-5535(03)00086-8</cr_doi><cr_issn type="print">15355535</cr_issn></crossref></journal-ref><journal-ref id="mst305869bib02"><authors><au><second-name>Lee</second-name><first-names>Y J</first-names></au><au><second-name>Chang</second-name><first-names>K H</first-names></au><au><second-name>Chon</second-name><first-names>J C</first-names></au><au><second-name>Oh</second-name><first-names>C Y</first-names></au></authors><year>2004</year><art-title>Automatic alignment method for calibration of hydrometers</art-title><jnl-title>Metrologia</jnl-title><volume>41</volume><pages>100–4</pages><crossref><cr_doi>http://dx.doi.org/10.1088/0026-1394/41/2/S11</cr_doi><cr_issn type="print">00261394</cr_issn><cr_issn type="electronic">16817575</cr_issn></crossref></journal-ref><journal-ref id="mst305869bib03"><authors><au><second-name>Hemming</second-name><first-names>B</first-names></au><au><second-name>Lehto</second-name><first-names>H</first-names></au></authors><year>2002</year><art-title>Calibration of dial indicators using machine vision</art-title><jnl-title>Meas. Sci. Technol.</jnl-title><volume>13</volume><pages>45–9</pages><crossref><cr_doi>http://dx.doi.org/10.1088/0957-0233/13/1/306</cr_doi><cr_issn type="print">09570233</cr_issn><cr_issn type="electronic">13616501</cr_issn></crossref></journal-ref><journal-ref id="mst305869bib04"><authors><au><second-name>Hemming</second-name><first-names>B</first-names></au><au><second-name>Fagerholm</second-name><first-names>A</first-names></au><au><second-name>Lassila</second-name><first-names>A</first-names></au></authors><year>2007</year><art-title>High-accuracy automatic machine vision based calibration of micrometers</art-title><jnl-title>Meas. Sci. Technol.</jnl-title><volume>18</volume><pages>1655–60</pages><crossref><cr_doi>http://dx.doi.org/10.1088/0957-0233/18/5/058</cr_doi><cr_issn type="print">09570233</cr_issn><cr_issn type="electronic">13616501</cr_issn></crossref></journal-ref><journal-ref id="mst305869bib05"><authors><au><second-name>Mori</second-name><first-names>S</first-names></au><au><second-name>Suen</second-name><first-names>C Y</first-names></au><au><second-name>Yamamoto</second-name><first-names>K</first-names></au></authors><year>1992</year><art-title>Historical review of OCR research and development</art-title><jnl-title>Proc. IEEE</jnl-title><volume>80</volume><pages>1029–58</pages><crossref><cr_doi>http://dx.doi.org/10.1109/5.156468</cr_doi><cr_issn type="print">00189219</cr_issn></crossref></journal-ref><journal-ref id="mst305869bib06"><authors><au><second-name>Correa Alegria</second-name><first-names>F</first-names></au><au><second-name>Cruz Serra</second-name><first-names>A</first-names></au></authors><year>2000</year><art-title>Automatic calibration of analog and digital measuring instruments using computer vision</art-title><jnl-title>IEEE Trans. Instrum. Meas.</jnl-title><volume>49</volume><pages>94–9</pages><crossref><cr_doi>http://dx.doi.org/10.1109/19.836317</cr_doi><cr_issn type="print">00189456</cr_issn></crossref></journal-ref><conf-ref id="mst305869bib07"><authors><au><second-name>Martín</second-name><first-names>F</first-names></au></authors><year>2003</year><art-title>Analysis tools for gray level histograms</art-title><conf-title>Proc. Int. Conf. on Signal Processing, Pattern Recognition and Applications</conf-title><editors><author><first-names>M H</first-names><second-name>Hamza</second-name></author></editors><publication><place>Rhodes, Greece</place><publisher>IASTED</publisher></publication></conf-ref><journal-ref id="mst305869bib08"><authors><au><second-name>Ohya</second-name><first-names>J</first-names></au><au><second-name>Shio</second-name><first-names>A</first-names></au><au><second-name>Akamatsu</second-name><first-names>S</first-names></au></authors><year>1994</year><art-title>Recognizing characters in scene images</art-title><jnl-title>IEEE Trans. Pattern Anal. Mach. Intell.</jnl-title><volume>16</volume><pages>214–20</pages><crossref><cr_doi>http://dx.doi.org/10.1109/34.273729</cr_doi><cr_issn type="print">01628828</cr_issn></crossref></journal-ref><journal-ref id="mst305869bib09"><authors><au><second-name>Otsu</second-name><first-names>N</first-names></au></authors><year>1979</year><art-title>A threshold selection method for gray level histograms</art-title><jnl-title>IEEE Trans. Syst. Man Cybern.</jnl-title><volume>9</volume><pages>62–6</pages><crossref><cr_doi>http://dx.doi.org/10.1109/TSMC.1979.4310076</cr_doi><cr_issn type="print">00189472</cr_issn></crossref></journal-ref><book-ref id="mst305869bib10"><authors><au><second-name>Gonzalez</second-name><first-names>R C</first-names></au><au><second-name>Woods</second-name><first-names>R E</first-names></au></authors><year>2008</year><book-title>Digital Image Processing</book-title><publication><place>Boston, MA</place><publisher>Prentice-Hall</publisher></publication></book-ref><journal-ref id="mst305869bib11"><authors><au><second-name>Lu</second-name><first-names>Y</first-names></au></authors><year>1995</year><art-title>Machine printed character segmentation-an overview</art-title><jnl-title>Pattern Recognit.</jnl-title><volume>28</volume><pages>67–80</pages><crossref><cr_doi>http://dx.doi.org/10.1016/0031-3203(94)00068-W</cr_doi><cr_issn type="print">00313203</cr_issn></crossref></journal-ref><journal-ref id="mst305869bib12"><authors><au><second-name>Kirsch</second-name><first-names>R A</first-names></au></authors><year>1971</year><art-title>Computer determination of the constituent structure of biological images</art-title><jnl-title>Comput. Biomed. Res.</jnl-title><volume>4</volume><pages>315–28</pages><crossref><cr_doi>http://dx.doi.org/10.1016/0010-4809(71)90034-6</cr_doi><cr_issn type="print">00104809</cr_issn><cr_issn type="electronic">10902368</cr_issn></crossref></journal-ref><conf-ref id="mst305869bib13"><authors><au><second-name>Vázquez-Fernández</second-name><first-names>E</first-names></au><au><second-name>González-Jorge</second-name><first-names>H</first-names></au><au><second-name>Dacal-Nieto</second-name><first-names>A</first-names></au><au><second-name>Martín</second-name><first-names>F</first-names></au><au><second-name>Formella</second-name><first-names>A</first-names></au></authors><year>2008</year><art-title>Human feature perception as a complementary method for digit recognition</art-title><conf-title>Proc. Int. Conf. on Visualization, Imaging, and Image Processing</conf-title><editors><author><first-names>J J</first-names><second-name>Villanueva</second-name></author></editors><publication><place>Palma de Mallorca, Spain</place><publisher>IASTED</publisher></publication><pages>pp 251–6</pages></conf-ref></reference-list></references></back></article></istex:document></istex:metadataXml><mods version="3.6"><titleInfo><title>A machine vision system for the calibration of digital thermometers</title></titleInfo><titleInfo type="abbreviated"><title>A machine vision system for the calibration of digital thermometers</title></titleInfo><titleInfo type="alternative"><title>A machine vision system for the calibration of digital thermometers</title></titleInfo><name type="personal"><namePart type="given">Esteban</namePart><namePart type="family">Vzquez-Fernndez</namePart><affiliation>Laboratorio Oficial de Metroloxa de Galicia (LOMG), Parque Tecnolxico de Galicia, San Cibrao das Vias, 32901, Ourense, Spain</affiliation><affiliation>E-mail:evazquez@lomg.net</affiliation><role><roleTerm type="text">author</roleTerm></role></name><name type="personal"><namePart type="given">Angel</namePart><namePart type="family">Dacal-Nieto</namePart><affiliation>Laboratorio Oficial de Metroloxa de Galicia (LOMG), Parque Tecnolxico de Galicia, San Cibrao das Vias, 32901, Ourense, Spain</affiliation><role><roleTerm type="text">author</roleTerm></role></name><name type="personal"><namePart type="given">Higinio</namePart><namePart type="family">Gonzlez-Jorge</namePart><affiliation>Laboratorio Oficial de Metroloxa de Galicia (LOMG), Parque Tecnolxico de Galicia, San Cibrao das Vias, 32901, Ourense, Spain</affiliation><role><roleTerm type="text">author</roleTerm></role></name><name type="personal"><namePart type="given">Fernando</namePart><namePart type="family">Martn</namePart><affiliation>Departamento de Teora do Sinal e Comunicacins, Universidade de Vigo, Campus Universitario, 36310, Vigo, Spain</affiliation><role><roleTerm type="text">author</roleTerm></role></name><name type="personal"><namePart type="given">Arno</namePart><namePart type="family">Formella</namePart><affiliation>Departamento de Informtica, Universidade de Vigo, As Lagoas s/n, 32004, Ourense, Spain</affiliation><role><roleTerm type="text">author</roleTerm></role></name><name type="personal"><namePart type="given">Victor</namePart><namePart type="family">Alvarez-Valado</namePart><affiliation>Laboratorio Oficial de Metroloxa de Galicia (LOMG), Parque Tecnolxico de Galicia, San Cibrao das Vias, 32901, Ourense, Spain</affiliation><role><roleTerm type="text">author</roleTerm></role></name><typeOfResource>text</typeOfResource><genre type="research-article" displayLabel="paper"></genre><originInfo><publisher>IOP Publishing</publisher><dateIssued encoding="w3cdtf">2009</dateIssued><copyrightDate encoding="w3cdtf">2009</copyrightDate></originInfo><language><languageTerm type="code" authority="iso639-2b">eng</languageTerm><languageTerm type="code" authority="rfc3066">en</languageTerm></language><physicalDescription><internetMediaType>text/html</internetMediaType><note type="production">Printed in the UK</note></physicalDescription><abstract>Automation is a key point in many industrial tasks such as calibration and metrology. In this context, machine vision has shown to be a useful tool for automation support, especially when there is no other option available. A system for the calibration of portable measurement devices has been developed. The system uses machine vision to obtain the numerical values shown by displays. A new approach based on human perception of digits, which works in parallel with other more classical classifiers, has been created. The results show the benefits of the system in terms of its usability and robustness, obtaining a success rate higher than 99 in display recognition. The system saves time and effort, and offers the possibility of scheduling calibration tasks without excessive attention by the laboratory technicians.</abstract><subject><genre>keywords</genre><topic>metrology</topic><topic>machine vision</topic><topic>calibration</topic><topic>thermometer</topic><topic>automation</topic></subject><classification authority="pacs">06.20.-f</classification><classification authority="pacs">06.20.fb</classification><classification authority="pacs">06.60.Mr</classification><classification authority="pacs">07.20.Dt</classification><relatedItem type="host"><titleInfo><title>Measurement Science and Technology</title></titleInfo><titleInfo type="abbreviated"><title>Meas. Sci. Technol.</title></titleInfo><genre type="journal">journal</genre><identifier type="ISSN">0957-0233</identifier><identifier type="eISSN">1361-6501</identifier><identifier type="PublisherID">MST</identifier><identifier type="CODEN">MSTCEP</identifier><identifier type="URL">stacks.iop.org/MST</identifier><part><date>2009</date><detail type="volume"><caption>vol.</caption><number>20</number></detail><detail type="issue"><caption>no.</caption><number>6</number></detail><extent unit="pages"><start>1</start><end>7</end><total>7</total></extent></part></relatedItem><identifier type="istex">EBF303D3EB21A4EED751C98AB5B05281A086F542</identifier><identifier type="DOI">10.1088/0957-0233/20/6/065106</identifier><identifier type="PII">S0957-0233(09)05869-X</identifier><identifier type="articleID">305869</identifier><identifier type="articleNumber">065106</identifier><accessCondition type="use and reproduction" contentType="copyright">2009 IOP Publishing Ltd</accessCondition><recordInfo><recordContentSource>IOP</recordContentSource><recordOrigin>2009 IOP Publishing Ltd</recordOrigin></recordInfo></mods></metadata><serie></serie></istex></record>Pour manipuler ce document sous Unix (Dilib)
EXPLOR_STEP=$WICRI_ROOT/Ticri/CIDE/explor/OcrV1/Data/Istex/Corpus
HfdSelect -h $EXPLOR_STEP/biblio.hfd -nk 000503 | SxmlIndent | more
Ou
HfdSelect -h $EXPLOR_AREA/Data/Istex/Corpus/biblio.hfd -nk 000503 | SxmlIndent | more
Pour mettre un lien sur cette page dans le réseau Wicri
{{Explor lien
|wiki= Ticri/CIDE
|area= OcrV1
|flux= Istex
|étape= Corpus
|type= RBID
|clé= ISTEX:EBF303D3EB21A4EED751C98AB5B05281A086F542
|texte= A machine vision system for the calibration of digital thermometers
}}
| This area was generated with Dilib version V0.6.32. |  |