IFAC Pape sOnLine 52-27 (2019) 508–513
ScienceDi ec
A ailable online a www.sciencedi ec .com
2405-8963 © 2019, IFAC (In e na ional Fede a ion o Au oma ic Con ol) Hos ing by Else ie L d. All igh s ese ed.
Pee e iew unde esponsibili y o In e na ional Fede a ion o Au oma ic Con ol.
10.1016/j.i acol.2019.12.714
©
2019, IFAC (In e na ional Fede a ion o Au oma ic Con ol) Hos ing by Else ie L d. All igh s ese ed.
10.1016/j.i acol.2019.12.714 2405-8963
T anspa en Laye Compensa ion a he Ac i e T iangula ion Sys ems
I. Janako a*, S. Sedi a**, P. Benes***
Depa men o Con ol and Ins umen a ion, Facul y o Elec ical
Enginee ing and Communica ion, B no Uni e si y o Technology,
B no, Czech Republic
*(+420 541 146 390, e-mail: janako a@ eec. u b .cz)
**(+420 541 146 418, e-mail: sedi a@ eec. u b .cz)
***(+420 541 141 152, e-mail: benes@ eec. u b .cz)
Abs ac : The pape deals wi h an ac i e iangula ion echnique o non-con ac op ical measu emen o
3D p ope ies. The basic ea u es o he me hod - p inciple, ad an ages and disad an ages - a e desc ibed
i s . Bu he main pa o he pape is conce ned wi h he impossibili y o using he basic p inciple o he
me hod o measu ed objec s, which a e co e ed by a laye o anspa en ma e ial o unknown hickness.
Mos o en i is a laye o oil o he p ese a ion o me al p oduc s. This laye dis o s he measu ed
alue due o he ligh e ac ion. In his wo k, a me hod o elimina ing his dis o ion based on he use o
wo senso s is p oposed. Expe imen al e i ica ion o his p oposal is also p esen ed.
Keywo ds: 3D ision, op ical dis ance measu emen , ac i e iangula ion, image p ocessing, anspa en
laye , oil compensa ion.
1. INTRODUCTION
Op ical non-con ac dis an measu ing echniques ha e ound
widesp ead use in many a eas, o example in a ic,
a chi ec u e, obo na iga ion, objec ecogni ion, o in
indus y in he p oduc ion inspec ion (Kalo á and Lisz wan,
2006). Thei expansion was suppo ed no only by he g ea
de elopmen o echnics (elec onics, op ics), bu also by he
p og ess made in echniques o da a and image p ocessing.
The ad an ages o op ical me hods a e ha hey can be e y
accu a e and as and agains o he p inciples ha a e non-
con ac and non-des uc i e. I is also easie o econ igu e
he sys em o ano he dimension o ype o p oduc s.
In s anda d p ojec i e imaging, one dimension (usually he
dep h in o ma ion) om 3D scenes is los . Wi hou he use o
a special echnique, i is no possible o de e mine his
in o ma ion e oac i ely. Fo una ely, depending on he
speci ic equi emen s, di e en echnical app oaches can be
used o op ical dis ance measu emen (Be ko ic and Sha i ,
2012). The mos used app oaches a e based on one o h ee
basic p inciples: iangula ion, ime-o - ligh (ToF) o op ical
in e e ome y (Schwa e e al., 1999).
Among he mos impo an pa ame e s o selec ing he mos
app op ia e me hod include he equi ed ange o measu ed
dis ances (dimensions o measu ed objec and he wo king
dis ance), esolu ion, measu ing ime, opology (conca i ies,
p o usions, edges) and su ace p ope ies ( oughness,
specula e lec ions) o he measu ed objec , p ope ies o
su ounding ligh ing, size o measu ing sys em, possibili ies
o calib a ion, and also cos s.
The mos widely used echnique o op ical o m
measu emen s is iangula ion. T iangles a e he basis o
many measu emen echniques. The ollowing a ian s,
which look g ea ly di e en , bu use same p inciple can be
dis inguished: ac i e iangula ion echniques wi h di e en
illumina ion, passi e iangula ion echniques on he basis o
digi al pho og amme y and s e eoscopy, heodoli e
measu ing sys ems, shape om shading echniques and ocus
echniques.
The ac i e iangula ion echnique is e y o en used in
p ac ice due o i s obus ness and e iciency, bu i also has
se e al limi a ions and disad an ages. In pa icula , his
a icle ocuses on he limi a ions o using his me hod o
objec s ha a e co e ed wi h a anspa en laye , such as
p ese ing oil.
2. ACTIVE TRIANGULATION
2.1 P inciple
The ac i e iangula ion echnique is based on a
pho og amme ic econs uc ion o he measu ed objec by
illumina ion i s su ace and con empo aneous scanning by a
senso (mos ly PSD module o CCD o CMOS came a). The
p inciple o his echnique is shown in Fig. 1. The ligh
sou ce, he de ec o and he illumina ed pa o he measu ed
objec o m a iangle. The join b be ween he ligh sou ce
and he de ec o is called a iangula ion op ical basis. The
ligh sou ce ay angle α is ixed whe eas he angle on he
de ec o side β changes depending on he a iable dis ance z
and i is de ined by an illumina ed poin on he came a chip,
espec i ely by a b igh poin on a p oduced image. I he
in es iga ed objec is si ua ed a he away om he came a
and ligh sou ce, he angle β is bigge so he e lec ed ligh
ay is p ojec ed o he pixel mo e up (acco ding o he
T anspa en Laye Compensa ion a he Ac i e T iangula ion Sys ems
I. Janako a*, S. Sedi a**, P. Benes***
Depa men o Con ol and Ins umen a ion, Facul y o Elec ical
Enginee ing and Communica ion, B no Uni e si y o Technology,
B no, Czech Republic
*(+420 541 146 390, e-mail: janako a@ eec. u b .cz)
**(+420 541 146 418, e-mail: sedi a@ eec. u b .cz)
***(+420 541 141 152, e-mail: benes@ eec. u b .cz)
Abs ac : The pape deals wi h an ac i e iangula ion echnique o non-con ac op ical measu emen o
3D p ope ies. The basic ea u es o he me hod - p inciple, ad an ages and disad an ages - a e desc ibed
i s . Bu he main pa o he pape is conce ned wi h he impossibili y o using he basic p inciple o he
me hod o measu ed objec s, which a e co e ed by a laye o anspa en ma e ial o unknown hickness.
Mos o en i is a laye o oil o he p ese a ion o me al p oduc s. This laye dis o s he measu ed
alue due o he ligh e ac ion. In his wo k, a me hod o elimina ing his dis o ion based on he use o
wo senso s is p oposed. Expe imen al e i ica ion o his p oposal is also p esen ed.
Keywo ds: 3D ision, op ical dis ance measu emen , ac i e iangula ion, image p ocessing, anspa en
laye , oil compensa ion.
1. INTRODUCTION
Op ical non-con ac dis an measu ing echniques ha e ound
widesp ead use in many a eas, o example in a ic,
a chi ec u e, obo na iga ion, objec ecogni ion, o in
indus y in he p oduc ion inspec ion (Kalo á and Lisz wan,
2006). Thei expansion was suppo ed no only by he g ea
de elopmen o echnics (elec onics, op ics), bu also by he
p og ess made in echniques o da a and image p ocessing.
The ad an ages o op ical me hods a e ha hey can be e y
accu a e and as and agains o he p inciples ha a e non-
con ac and non-des uc i e. I is also easie o econ igu e
he sys em o ano he dimension o ype o p oduc s.
In s anda d p ojec i e imaging, one dimension (usually he
dep h in o ma ion) om 3D scenes is los . Wi hou he use o
a special echnique, i is no possible o de e mine his
in o ma ion e oac i ely. Fo una ely, depending on he
speci ic equi emen s, di e en echnical app oaches can be
used o op ical dis ance measu emen (Be ko ic and Sha i ,
2012). The mos used app oaches a e based on one o h ee
basic p inciples: iangula ion, ime-o - ligh (ToF) o op ical
in e e ome y (Schwa e e al., 1999).
Among he mos impo an pa ame e s o selec ing he mos
app op ia e me hod include he equi ed ange o measu ed
dis ances (dimensions o measu ed objec and he wo king
dis ance), esolu ion, measu ing ime, opology (conca i ies,
p o usions, edges) and su ace p ope ies ( oughness,
specula e lec ions) o he measu ed objec , p ope ies o
su ounding ligh ing, size o measu ing sys em, possibili ies
o calib a ion, and also cos s.
The mos widely used echnique o op ical o m
measu emen s is iangula ion. T iangles a e he basis o
many measu emen echniques. The ollowing a ian s,
which look g ea ly di e en , bu use same p inciple can be
dis inguished: ac i e iangula ion echniques wi h di e en
illumina ion, passi e iangula ion echniques on he basis o
digi al pho og amme y and s e eoscopy, heodoli e
measu ing sys ems, shape om shading echniques and ocus
echniques.
The ac i e iangula ion echnique is e y o en used in
p ac ice due o i s obus ness and e iciency, bu i also has
se e al limi a ions and disad an ages. In pa icula , his
a icle ocuses on he limi a ions o using his me hod o
objec s ha a e co e ed wi h a anspa en laye , such as
p ese ing oil.
2. ACTIVE TRIANGULATION
2.1 P inciple
The ac i e iangula ion echnique is based on a
pho og amme ic econs uc ion o he measu ed objec by
illumina ion i s su ace and con empo aneous scanning by a
senso (mos ly PSD module o CCD o CMOS came a). The
p inciple o his echnique is shown in Fig. 1. The ligh
sou ce, he de ec o and he illumina ed pa o he measu ed
objec o m a iangle. The join b be ween he ligh sou ce
and he de ec o is called a iangula ion op ical basis. The
ligh sou ce ay angle α is ixed whe eas he angle on he
de ec o side β changes depending on he a iable dis ance z
and i is de ined by an illumina ed poin on he came a chip,
espec i ely by a b igh poin on a p oduced image. I he
in es iga ed objec is si ua ed a he away om he came a
and ligh sou ce, he angle β is bigge so he e lec ed ligh
ay is p ojec ed o he pixel mo e up (acco ding o he
T anspa en Laye Compensa ion a he Ac i e T iangula ion Sys ems
I. Janako a*, S. Sedi a**, P. Benes***
Depa men o Con ol and Ins umen a ion, Facul y o Elec ical
Enginee ing and Communica ion, B no Uni e si y o Technology,
B no, Czech Republic
*(+420 541 146 390, e-mail: janako a@ eec. u b .cz)
**(+420 541 146 418, e-mail: sedi a@ eec. u b .cz)
***(+420 541 141 152, e-mail: benes@ eec. u b .cz)
Abs ac : The pape deals wi h an ac i e iangula ion echnique o non-con ac op ical measu emen o
3D p ope ies. The basic ea u es o he me hod - p inciple, ad an ages and disad an ages - a e desc ibed
i s . Bu he main pa o he pape is conce ned wi h he impossibili y o using he basic p inciple o he
me hod o measu ed objec s, which a e co e ed by a laye o anspa en ma e ial o unknown hickness.
Mos o en i is a laye o oil o he p ese a ion o me al p oduc s. This laye dis o s he measu ed
alue due o he ligh e ac ion. In his wo k, a me hod o elimina ing his dis o ion based on he use o
wo senso s is p oposed. Expe imen al e i ica ion o his p oposal is also p esen ed.
Keywo ds: 3D ision, op ical dis ance measu emen , ac i e iangula ion, image p ocessing, anspa en
laye , oil compensa ion.
1. INTRODUCTION
Op ical non-con ac dis an measu ing echniques ha e ound
widesp ead use in many a eas, o example in a ic,
a chi ec u e, obo na iga ion, objec ecogni ion, o in
indus y in he p oduc ion inspec ion (Kalo á and Lisz wan,
2006). Thei expansion was suppo ed no only by he g ea
de elopmen o echnics (elec onics, op ics), bu also by he
p og ess made in echniques o da a and image p ocessing.
The ad an ages o op ical me hods a e ha hey can be e y
accu a e and as and agains o he p inciples ha a e non-
con ac and non-des uc i e. I is also easie o econ igu e
he sys em o ano he dimension o ype o p oduc s.
In s anda d p ojec i e imaging, one dimension (usually he
dep h in o ma ion) om 3D scenes is los . Wi hou he use o
a special echnique, i is no possible o de e mine his
in o ma ion e oac i ely. Fo una ely, depending on he
speci ic equi emen s, di e en echnical app oaches can be
used o op ical dis ance measu emen (Be ko ic and Sha i ,
2012). The mos used app oaches a e based on one o h ee
basic p inciples: iangula ion, ime-o - ligh (ToF) o op ical
in e e ome y (Schwa e e al., 1999).
Among he mos impo an pa ame e s o selec ing he mos
app op ia e me hod include he equi ed ange o measu ed
dis ances (dimensions o measu ed objec and he wo king
dis ance), esolu ion, measu ing ime, opology (conca i ies,
p o usions, edges) and su ace p ope ies ( oughness,
specula e lec ions) o he measu ed objec , p ope ies o
su ounding ligh ing, size o measu ing sys em, possibili ies
o calib a ion, and also cos s.
The mos widely used echnique o op ical o m
measu emen s is iangula ion. T iangles a e he basis o
many measu emen echniques. The ollowing a ian s,
which look g ea ly di e en , bu use same p inciple can be
dis inguished: ac i e iangula ion echniques wi h di e en
illumina ion, passi e iangula ion echniques on he basis o
digi al pho og amme y and s e eoscopy, heodoli e
measu ing sys ems, shape om shading echniques and ocus
echniques.
The ac i e iangula ion echnique is e y o en used in
p ac ice due o i s obus ness and e iciency, bu i also has
se e al limi a ions and disad an ages. In pa icula , his
a icle ocuses on he limi a ions o using his me hod o
objec s ha a e co e ed wi h a anspa en laye , such as
p ese ing oil.
2. ACTIVE TRIANGULATION
2.1 P inciple
The ac i e iangula ion echnique is based on a
pho og amme ic econs uc ion o he measu ed objec by
illumina ion i s su ace and con empo aneous scanning by a
senso (mos ly PSD module o CCD o CMOS came a). The
p inciple o his echnique is shown in Fig. 1. The ligh
sou ce, he de ec o and he illumina ed pa o he measu ed
objec o m a iangle. The join b be ween he ligh sou ce
and he de ec o is called a iangula ion op ical basis. The
ligh sou ce ay angle α is ixed whe eas he angle on he
de ec o side β changes depending on he a iable dis ance z
and i is de ined by an illumina ed poin on he came a chip,
espec i ely by a b igh poin on a p oduced image. I he
in es iga ed objec is si ua ed a he away om he came a
and ligh sou ce, he angle β is bigge so he e lec ed ligh
ay is p ojec ed o he pixel mo e up (acco ding o he
T anspa en Laye Compensa ion a he Ac i e T iangula ion Sys ems
I. Janako a*, S. Sedi a**, P. Benes***
Depa men o Con ol and Ins umen a ion, Facul y o Elec ical
Enginee ing and Communica ion, B no Uni e si y o Technology,
B no, Czech Republic
*(+420 541 146 390, e-mail: janako a@ eec. u b .cz)
**(+420 541 146 418, e-mail: sedi a@ eec. u b .cz)
***(+420 541 141 152, e-mail: benes@ eec. u b .cz)
Abs ac : The pape deals wi h an ac i e iangula ion echnique o non-con ac op ical measu emen o
3D p ope ies. The basic ea u es o he me hod - p inciple, ad an ages and disad an ages - a e desc ibed
i s . Bu he main pa o he pape is conce ned wi h he impossibili y o using he basic p inciple o he
me hod o measu ed objec s, which a e co e ed by a laye o anspa en ma e ial o unknown hickness.
Mos o en i is a laye o oil o he p ese a ion o me al p oduc s. This laye dis o s he measu ed
alue due o he ligh e ac ion. In his wo k, a me hod o elimina ing his dis o ion based on he use o
wo senso s is p oposed. Expe imen al e i ica ion o his p oposal is also p esen ed.
Keywo ds: 3D ision, op ical dis ance measu emen , ac i e iangula ion, image p ocessing, anspa en
laye , oil compensa ion.
1. INTRODUCTION
Op ical non-con ac dis an measu ing echniques ha e ound
widesp ead use in many a eas, o example in a ic,
a chi ec u e, obo na iga ion, objec ecogni ion, o in
indus y in he p oduc ion inspec ion (Kalo á and Lisz wan,
2006). Thei expansion was suppo ed no only by he g ea
de elopmen o echnics (elec onics, op ics), bu also by he
p og ess made in echniques o da a and image p ocessing.
The ad an ages o op ical me hods a e ha hey can be e y
accu a e and as and agains o he p inciples ha a e non-
con ac and non-des uc i e. I is also easie o econ igu e
he sys em o ano he dimension o ype o p oduc s.
In s anda d p ojec i e imaging, one dimension (usually he
dep h in o ma ion) om 3D scenes is los . Wi hou he use o
a special echnique, i is no possible o de e mine his
in o ma ion e oac i ely. Fo una ely, depending on he
speci ic equi emen s, di e en echnical app oaches can be
used o op ical dis ance measu emen (Be ko ic and Sha i ,
2012). The mos used app oaches a e based on one o h ee
basic p inciples: iangula ion, ime-o - ligh (ToF) o op ical
in e e ome y (Schwa e e al., 1999).
Among he mos impo an pa ame e s o selec ing he mos
app op ia e me hod include he equi ed ange o measu ed
dis ances (dimensions o measu ed objec and he wo king
dis ance), esolu ion, measu ing ime, opology (conca i ies,
p o usions, edges) and su ace p ope ies ( oughness,
specula e lec ions) o he measu ed objec , p ope ies o
su ounding ligh ing, size o measu ing sys em, possibili ies
o calib a ion, and also cos s.
The mos widely used echnique o op ical o m
measu emen s is iangula ion. T iangles a e he basis o
many measu emen echniques. The ollowing a ian s,
which look g ea ly di e en , bu use same p inciple can be
dis inguished: ac i e iangula ion echniques wi h di e en
illumina ion, passi e iangula ion echniques on he basis o
digi al pho og amme y and s e eoscopy, heodoli e
measu ing sys ems, shape om shading echniques and ocus
echniques.
The ac i e iangula ion echnique is e y o en used in
p ac ice due o i s obus ness and e iciency, bu i also has
se e al limi a ions and disad an ages. In pa icula , his
a icle ocuses on he limi a ions o using his me hod o
objec s ha a e co e ed wi h a anspa en laye , such as
p ese ing oil.
2. ACTIVE TRIANGULATION
2.1 P inciple
The ac i e iangula ion echnique is based on a
pho og amme ic econs uc ion o he measu ed objec by
illumina ion i s su ace and con empo aneous scanning by a
senso (mos ly PSD module o CCD o CMOS came a). The
p inciple o his echnique is shown in Fig. 1. The ligh
sou ce, he de ec o and he illumina ed pa o he measu ed
objec o m a iangle. The join b be ween he ligh sou ce
and he de ec o is called a iangula ion op ical basis. The
ligh sou ce ay angle α is ixed whe eas he angle on he
de ec o side β changes depending on he a iable dis ance z
and i is de ined by an illumina ed poin on he came a chip,
espec i ely by a b igh poin on a p oduced image. I he
in es iga ed objec is si ua ed a he away om he came a
and ligh sou ce, he angle β is bigge so he e lec ed ligh
ay is p ojec ed o he pixel mo e up (acco ding o he
T anspa en Laye Compensa ion a he Ac i e T iangula ion Sys ems
I. Janako a*, S. Sedi a**, P. Benes***
Depa men o Con ol and Ins umen a ion, Facul y o Elec ical
Enginee ing and Communica ion, B no Uni e si y o Technology,
B no, Czech Republic
*(+420 541 146 390, e-mail: janako a@ eec. u b .cz)
**(+420 541 146 418, e-mail: sedi a@ eec. u b .cz)
***(+420 541 141 152, e-mail: benes@ eec. u b .cz)
Abs ac : The pape deals wi h an ac i e iangula ion echnique o non-con ac op ical measu emen o
3D p ope ies. The basic ea u es o he me hod - p inciple, ad an ages and disad an ages - a e desc ibed
i s . Bu he main pa o he pape is conce ned wi h he impossibili y o using he basic p inciple o he
me hod o measu ed objec s, which a e co e ed by a laye o anspa en ma e ial o unknown hickness.
Mos o en i is a laye o oil o he p ese a ion o me al p oduc s. This laye dis o s he measu ed
alue due o he ligh e ac ion. In his wo k, a me hod o elimina ing his dis o ion based on he use o
wo senso s is p oposed. Expe imen al e i ica ion o his p oposal is also p esen ed.
Keywo ds: 3D ision, op ical dis ance measu emen , ac i e iangula ion, image p ocessing, anspa en
laye , oil compensa ion.
1. INTRODUCTION
Op ical non-con ac dis an measu ing echniques ha e ound
widesp ead use in many a eas, o example in a ic,
a chi ec u e, obo na iga ion, objec ecogni ion, o in
indus y in he p oduc ion inspec ion (Kalo á and Lisz wan,
2006). Thei expansion was suppo ed no only by he g ea
de elopmen o echnics (elec onics, op ics), bu also by he
p og ess made in echniques o da a and image p ocessing.
The ad an ages o op ical me hods a e ha hey can be e y
accu a e and as and agains o he p inciples ha a e non-
con ac and non-des uc i e. I is also easie o econ igu e
he sys em o ano he dimension o ype o p oduc s.
In s anda d p ojec i e imaging, one dimension (usually he
dep h in o ma ion) om 3D scenes is los . Wi hou he use o
a special echnique, i is no possible o de e mine his
in o ma ion e oac i ely. Fo una ely, depending on he
speci ic equi emen s, di e en echnical app oaches can be
used o op ical dis ance measu emen (Be ko ic and Sha i ,
2012). The mos used app oaches a e based on one o h ee
basic p inciples: iangula ion, ime-o - ligh (ToF) o op ical
in e e ome y (Schwa e e al., 1999).
Among he mos impo an pa ame e s o selec ing he mos
app op ia e me hod include he equi ed ange o measu ed
dis ances (dimensions o measu ed objec and he wo king
dis ance), esolu ion, measu ing ime, opology (conca i ies,
p o usions, edges) and su ace p ope ies ( oughness,
specula e lec ions) o he measu ed objec , p ope ies o
su ounding ligh ing, size o measu ing sys em, possibili ies
o calib a ion, and also cos s.
The mos widely used echnique o op ical o m
measu emen s is iangula ion. T iangles a e he basis o
many measu emen echniques. The ollowing a ian s,
which look g ea ly di e en , bu use same p inciple can be
dis inguished: ac i e iangula ion echniques wi h di e en
illumina ion, passi e iangula ion echniques on he basis o
digi al pho og amme y and s e eoscopy, heodoli e
measu ing sys ems, shape om shading echniques and ocus
echniques.
The ac i e iangula ion echnique is e y o en used in
p ac ice due o i s obus ness and e iciency, bu i also has
se e al limi a ions and disad an ages. In pa icula , his
a icle ocuses on he limi a ions o using his me hod o
objec s ha a e co e ed wi h a anspa en laye , such as
p ese ing oil.
2. ACTIVE TRIANGULATION
2.1 P inciple
The ac i e iangula ion echnique is based on a
pho og amme ic econs uc ion o he measu ed objec by
illumina ion i s su ace and con empo aneous scanning by a
senso (mos ly PSD module o CCD o CMOS came a). The
p inciple o his echnique is shown in Fig. 1. The ligh
sou ce, he de ec o and he illumina ed pa o he measu ed
objec o m a iangle. The join b be ween he ligh sou ce
and he de ec o is called a iangula ion op ical basis. The
ligh sou ce ay angle α is ixed whe eas he angle on he
de ec o side β changes depending on he a iable dis ance z
and i is de ined by an illumina ed poin on he came a chip,
espec i ely by a b igh poin on a p oduced image. I he
in es iga ed objec is si ua ed a he away om he came a
and ligh sou ce, he angle β is bigge so he e lec ed ligh
ay is p ojec ed o he pixel mo e up (acco ding o he
T anspa en Laye Compensa ion a he Ac i e T iangula ion Sys ems
I. Janako a*, S. Sedi a**, P. Benes***
Depa men o Con ol and Ins umen a ion, Facul y o Elec ical
Enginee ing and Communica ion, B no Uni e si y o Technology,
B no, Czech Republic
*(+420 541 146 390, e-mail: janako a@ eec. u b .cz)
**(+420 541 146 418, e-mail: sedi a@ eec. u b .cz)
***(+420 541 141 152, e-mail: benes@ eec. u b .cz)
Abs ac : The pape deals wi h an ac i e iangula ion echnique o non-con ac op ical measu emen o
3D p ope ies. The basic ea u es o he me hod - p inciple, ad an ages and disad an ages - a e desc ibed
i s . Bu he main pa o he pape is conce ned wi h he impossibili y o using he basic p inciple o he
me hod o measu ed objec s, which a e co e ed by a laye o anspa en ma e ial o unknown hickness.
Mos o en i is a laye o oil o he p ese a ion o me al p oduc s. This laye dis o s he measu ed
alue due o he ligh e ac ion. In his wo k, a me hod o elimina ing his dis o ion based on he use o
wo senso s is p oposed. Expe imen al e i ica ion o his p oposal is also p esen ed.
Keywo ds: 3D ision, op ical dis ance measu emen , ac i e iangula ion, image p ocessing, anspa en
laye , oil compensa ion.
1. INTRODUCTION
Op ical non-con ac dis an measu ing echniques ha e ound
widesp ead use in many a eas, o example in a ic,
a chi ec u e, obo na iga ion, objec ecogni ion, o in
indus y in he p oduc ion inspec ion (Kalo á and Lisz wan,
2006). Thei expansion was suppo ed no only by he g ea
de elopmen o echnics (elec onics, op ics), bu also by he
p og ess made in echniques o da a and image p ocessing.
The ad an ages o op ical me hods a e ha hey can be e y
accu a e and as and agains o he p inciples ha a e non-
con ac and non-des uc i e. I is also easie o econ igu e
he sys em o ano he dimension o ype o p oduc s.
In s anda d p ojec i e imaging, one dimension (usually he
dep h in o ma ion) om 3D scenes is los . Wi hou he use o
a special echnique, i is no possible o de e mine his
in o ma ion e oac i ely. Fo una ely, depending on he
speci ic equi emen s, di e en echnical app oaches can be
used o op ical dis ance measu emen (Be ko ic and Sha i ,
2012). The mos used app oaches a e based on one o h ee
basic p inciples: iangula ion, ime-o - ligh (ToF) o op ical
in e e ome y (Schwa e e al., 1999).
Among he mos impo an pa ame e s o selec ing he mos
app op ia e me hod include he equi ed ange o measu ed
dis ances (dimensions o measu ed objec and he wo king
dis ance), esolu ion, measu ing ime, opology (conca i ies,
p o usions, edges) and su ace p ope ies ( oughness,
specula e lec ions) o he measu ed objec , p ope ies o
su ounding ligh ing, size o measu ing sys em, possibili ies
o calib a ion, and also cos s.
The mos widely used echnique o op ical o m
measu emen s is iangula ion. T iangles a e he basis o
many measu emen echniques. The ollowing a ian s,
which look g ea ly di e en , bu use same p inciple can be
dis inguished: ac i e iangula ion echniques wi h di e en
illumina ion, passi e iangula ion echniques on he basis o
digi al pho og amme y and s e eoscopy, heodoli e
measu ing sys ems, shape om shading echniques and ocus
echniques.
The ac i e iangula ion echnique is e y o en used in
p ac ice due o i s obus ness and e iciency, bu i also has
se e al limi a ions and disad an ages. In pa icula , his
a icle ocuses on he limi a ions o using his me hod o
objec s ha a e co e ed wi h a anspa en laye , such as
p ese ing oil.
2. ACTIVE TRIANGULATION
2.1 P inciple
The ac i e iangula ion echnique is based on a
pho og amme ic econs uc ion o he measu ed objec by
illumina ion i s su ace and con empo aneous scanning by a
senso (mos ly PSD module o CCD o CMOS came a). The
p inciple o his echnique is shown in Fig. 1. The ligh
sou ce, he de ec o and he illumina ed pa o he measu ed
objec o m a iangle. The join b be ween he ligh sou ce
and he de ec o is called a iangula ion op ical basis. The
ligh sou ce ay angle α is ixed whe eas he angle on he
de ec o side β changes depending on he a iable dis ance z
and i is de ined by an illumina ed poin on he came a chip,
espec i ely by a b igh poin on a p oduced image. I he
in es iga ed objec is si ua ed a he away om he came a
and ligh sou ce, he angle β is bigge so he e lec ed ligh
ay is p ojec ed o he pixel mo e up (acco ding o he
I. Janako a e al. / IFAC Pape sOnLine 52-27 (2019) 508–513 509
simpli ied Fig. 1). In his way, he dis ance o he objec can
be ob ained jus om he posi ion o he ligh poin in he
image. Howe e , wi h he α angle o he han 90 °, i should
be bo ne in mind ha when he objec 's dis ance changes, he
x posi ion on he su ace being measu ed also changes. The
ela ionship be ween he coo dina e in he image and he
dis ance can be de e mined by calcula ion (based on he
knowledge o he geome y and pa ame e s o he came a and
lens) o due o calib a ion.
Fig. 1. P inciple o he iangula ion me hod (1D a ian )
The measu emen unce ain y in z is in e sely p opo ional o
bo h he came a baseline and he ocal leng h o he lens, bu
di ec ly p opo ional o he squa e o he dis ance (Kalo á
and Lisz wan, 2005). Un o una ely, and b canno be made
as la ge as desi ed. The baseline is limi ed mainly by he
mechanical cons uc ion (s abili y o he whole sys em
dec eases as b inc eases), and by shadow e ec s (sel -
occlusion p oblems inc ease wi h b) (Be aldin
e al., 2003).
The size o he ocal leng h is limi ed by he limi ed
dimensions o he o e ed de ec o s. Mo eo e , wi h
inc easing he de ec o 's ield o iew dec eases and hus he
ange in he Z-axis also dec eases.
2.2 Va ian s
Depending on he ligh sou ce used o ma k he objec , h ee
a ian s can be dis inguished: a ligh poin , a ligh s ipe, and
a s uc u ed ligh . The dimension o he used ligh de e mines
he dimension and he numbe o ob ained poin s. When
using he ligh poin , only one coo dina e ( ypically dis ance,
dep h, o z coo dina e) o one poin is de e mined a a ime,
he e o e his a ian is called 1D iangula ion. The
ad an age o his op ion is ha a line came a wi h a usually
highe ho izon al esolu ion and high line scan equency can
be used. When he ligh s ipe (2D iangula ion) is used, wo
coo dina es a e de e mined a once and he numbe o
measu ed poin s depends on he came a's pixel esolu ion (a
p o ile is measu ed). The ad an age o using s uc u ed ligh -
3D iangula ion (Geng, 2011) is ha he comple e 3D
coo dina es o he whole objec can be de e mined in block
( he numbe o poin s depends on he used pa e n and he
came a esolu ion; he dimensions o he measu ed pa o he
objec a e de ined by he selec ed chip and lens and o e all
geome y). I i is equi ed o speci y mo e coo dina es ( o
1D and 2D a ian s) o mo e poin s, i is necessa y o ensu e
scanning (mo emen o he measu ing sys em o objec in
one o wo axes). Con en ional ligh sou ces can be used o
op ical iangula ion, bu lase sou ces ha e unique
ad an ages o 3D imaging (Amann e al., 2001). Abo e all,
i is high b igh ness, cohe ence and monoch oma ism, and
low di e gence.
2.3 Ad an ages and disad an ages
Ac i e iangula ion sys ems a e ela i ely simple, obus and
e icien . The ad an age is ha hey can ope a e o e a wide
ange o dis ances. Lase -based op ical displacemen senso s
using a e y small spo which enables measu emen s on he
e y small pa s. The 3D e sion can measu e an en i e objec
in one ins an om one ame. The 2D a ian is o en used
on p oduc ion lines, whe e scanning and measu emen o he
whole objec a e ensu ed hanks o he ansla ional
mo emen below he measu ing uni . The a ian wi h poin
lase and line came a is e y as , wi h a sampling a e o
100 kHz commonly achie ed. High measu ing a es a e
equi ed o he as mo ing a ge s o measu emen s on
di icul su aces. Comme cial lase scanne s a e e y obus
wi h compac dimensions. Despi e hei e y small
dimensions, some se ies ha e a ully in eg a ed con olle . As
a esul , simple, apid ins alla ion and wi ing a e possible.
The senso s can be in eg a ed easily e en in o es ic ed
ins alla ion space. The comp ehensi e solu ion enables easy
measu emen con ol, synch oniza ion wi h o he elemen s as
well as be ween se e al iangula ion senso s (e.g. di e en ial
hickness measu emen ) and usually a clea con ol and
e alua ion applica ion. The senso s can be also equipped wi h
di e en ou pu s o ul il indus ial use equi emen s.
The pi alls o ac i e iangula ion echniques a e gene ally o
e alua e pa s o he objec ha a e no di ec ly isible by he
came a o illumina ed by he ligh sou ce. Su ace opology
and geome y o eal physical objec s can be qui e complex,
wi h mul iple holes, conca i ies, p o usions and edges.
P oblems also b ing su aces pe pendicula o he image
plane and su aces pa allel o he ligh sou ce. The
measu emen quali y is also in luenced by he su ace
p ope ies o he measu ed objec . Essen ial a e especially
su ace oughness, e lec ance, and colou . Howe e , oday's
sys ems a e capable o measu ing e en complica ed su aces,
such as shiny me allic ma e ials o con a iwise ubbe . The
exposu e ime o he amoun o ligh p oduced by he lase is
op imally ma ched o he e lec ion cha ac e is ics o he
a ge su ace. New possibili ies also include he use o a blue
lase (nex o mo e common ed). Due o i s sho e
wa eleng h, i does no pene a e he a ge su ace,
p ojec ing a small ligh spo on he su ace and he e o e
p o iding s able and p ecise esul s. This echnology is
p e e ably used wi h o ganic and (semi-) anspa en objec s,
as well as o ed-ho glowing me als.
Ano he p oblem, ha was encoun e ed when es ing he lase
sys em o check he dimensions o p oduc s, can be caused by
he hin laye o some anspa en ma e ial on he p oduc s,
whe e he measu ed alues a e dis o ed due o he e ac ion
o he ligh in ha laye .
510 I. Janako a e al. / IFAC Pape sOnLine 52-27 (2019) 508–513
3. PROBLEM DEFINITION
The ask o he p oposed sys em was o measu e he
pa ame e s o he pa icula me al pa s o he axial bea ings
and, based on he measu emen s, o check whe he some pa s
we en' con used. Indi idual ypes o bea ings di e in he
combina ion o hei dimensions (heigh , inne and ou e
diame e , aceway pa ame e s) e en in he o de o only a
ew mic ome e s, so i was necessa y o ensu e e y accu a e
measu emen s. Un o una ely, he con ol measu ing sys em
had o be placed un il jus be o e assembling all pa s, whe e
he pa s a e oil-p ese ed. Due o he me hod o p ese a ion
by dipping and ee d ipping o he componen s in an
unde ined posi ion, he laye hickness is no uni o m and
canno be p ede e mined.
Se e al app oaches we e conside ed o measu emen . One
op ion was o use he 1D o 2D lase iangula ion sys em.
Howe e , he p oblem is ha he pa ially anspa en laye o
p ese a i e oil causes no only dis u bing ligh e lec ions
om i s su ace and also educes he in ensi y o ecei ed
ligh , bu abo e all, i causes a change in he di ec ion o he
ligh beam a he in e ace be ween wo media wi h di e en
e ac i e indices. I he lase beam passes om ambien ai
o an op ically dense oil medium ( he e ac i e index o he
oil is abou 1.5), he angle o e ac ion is less han he angle
o incidence - he ay in he highe -index medium is close o
he no mal. Respec i ely, when eco ding he ligh e lec ed
om he me al su ace, he opposi e phenomenon is
obse ed. Fo he sake o cla i y, he p oblem is shown in
Fig. 2 in a simple and mo e a ou able si ua ion, whe e he
lase beam is pe pendicula o he su ace ( he e o e he e is
no double e ac ion) and pa allel p ojec ion is used.
Fig. 2. The dis ance measu ing p oblem o componen s wi h
he oil laye
Poin A in Fig. 2 indica es he loca ion whe e he ligh beam
would be p ojec ed i he e we e no oil laye s. Con e sely,
poin B ep esen s he illumina ed spo on he chip a e
e ac ion o he ligh in he uppe anspa en laye . Due o
he ac ha he p ojec ion is shi ed on he chip, he w ong
dis ance l
B
is de e mined agains he eal dis ance l o he
me al su ace o he bea ing om he lase .
The magni ude o he e o is di ec ly p opo ional o he
angle be ween he came a's op ical axis and he no mal o he
measu ed su ace α, as well as o he e ac i e index o he
laye n
2
and o i s hickness . T igonome ic unc ions can be
used o calcula e he e o ( o simplici y, a elecen ic lens
wi h pa allel beams is conside ed, as in Fig. 2):
α
β
g
g e
⋅
−=
, (1)
whe e angle
β is gi en by Snell’s law:
βα
sinsin
21
⋅=⋅ nn
, (2)
Fo example, o he angle α = 50° and he oil laye (
n
2
= 1.5)
wi h he hickness o only 0.1 mm, he e o is g ea e han
0.05 mm, so o he co ec dis inc ion o bea ing ypes is his
e o e y signi ican . The hickness o he oil laye may a y
in di e en pa s o he bea ing ing due o he su ace ension
and unning down oil in o he bea ing aceway (see Fig. 2),
and hus depends also on he size and shape o he ing. In
addi ion, he oil hickness can be changed by u he
manipula ion, o example by g ipping he ing in he hand.
Howe e , due o he unknown ac ual oil hickness, i is no
possible o de e mine he amoun o dis o ion, espec i ely
enqui ed bea ing heigh , om only one measu emen .
4. SUGGESTED SOLUTION
The p oposed solu ion o he p oblem desc ibed in he
p e ious chap e is o use wo senso s ( wo came as) each
wi h he di e en angle o he lase beam. The p inciple is
shown in Fig. 3.
Fig. 3. P inciple o he p oposed solu ion using wo senso s
Each sepa a e measu ing uni , composed o he came a and
he lase (whe e he lase can be common o bo h uni s)
measu es, due o di e en iewing angle α
i
, he di e sely
dis o ed coo dina es
z
ei
(
i
is he index o measu emen )
iiii iii
i
e
ezz nn z +== ), ,,,(
21
α
. (3)
I. Janako a e al. / IFAC Pape sOnLine 52-27 (2019) 508–513 511
The medium
n
1
, whe e he whole measu ing uni is loca ed, is
su ounding ai so
n
11
=
n
12
=
n
1
= 1
. The medium
n
2
ep esen s he anspa en laye abo e he measu ed objec (in
ou case oil, whe e
n
2
is app oxima ely 1.5) and i is uni o m
o bo h measu emen s (
n
21
=
n
22
=
n
2
). In he nea
neighbou hood o he measu emen , unknown hicknesses o
he laye can also be conside ed iden ical (
1
=
2
=
). And o
cou se, equal is he second unknown in ou ask - he heigh
o he objec abo e he pad
z
1
=
z
2
=
z
. I he angles α
i
and
index
n
2
a e known, wo equa ions o wo unknowns can be
cons uc ed based on wo measu emen s and
z
and
can be
de e mined simul aneously. The ad an age is ha he same
sys em can be used o ano he ype o ask wi h wo
unknowns. Fo example, he laye e ac i e index o he
unknown hickness can be de e mined o ice e sa.
5. EXPERIMENTAL VERIFICATION
A labo a o y iangula ion wo kplace wi h wo came as was
assembled o e i y he p oposed solu ion. The design was
es ed on he es ask o measu ing he pa ame e s o axial
bea ings. Fo he measu emen , he wo-dimensional a ian
wi h he line lase was chosen o measu e he en i e bea ing
p o ile a once.
The a angemen o he indi idual componen s o he
assembled expe imen al wo kplace is shown in he
pho og aph (Fig. 4) and he ske ch (Fig.5). The bea ing was
placed on a la backpla e wi h he aceway poin ing upwa ds
and was pushed o wo chocks (diame e 6 mm and dis ance
25 mm). This ensu es p ecise posi ioning, cen e ing o he
ing and, a he same ime, he ou e diame e o he ing can
be de e mined om he size o inse ion o he ing be ween
he s ops (p o ile posi ion wi hin he measu ed ield). The
line lase was moun ed abo e he pad o shine pe pendicula
o he pad, and he lase s ipe illumina ed he space exac ly
be ween he s oppe s. The op ical plane o he lase hen
passes h ough he symme y axis o he ing. The came as
we e moun ed sepa a ely on he s ands and we e o ien ed so
ha he lase ack e lec ed om he pad was a he bo om
o he cap u ed images and was pa allel o he Y-axis o he
images and, o cou se, o wa ch he a ea a he ing p o ile.
Fig. 4. Pho og aph o he assembled expe imen al wo kplace
Fig. 5. Schema ic ske ch o he a angemen o he indi idual
componen s o he designed measu ing sys em
Two iden ical CCD came as we e used (Imaging Sou ce
DFK 41BU02 wi h 1/2" chip size and 1280x960 pixel
esolu ion). A na owband il e was inse ed in on o each
chip o il e ou ligh componen s o he han he wa eleng h
o he lase used (650 nm). Bo h came as we e supplemen ed
wi h he same lenses (VICO imaging MFA1-230-5M35 wi h
35 mm ocal leng h and 1:1.4 ape u e). A dis ance ing has
always been inse ed be ween he came a and he lens o
educe he wo king dis ance. The came a dis ance was
adjus ed o ob ain he necessa y X-axis ange (app oxima ely
26 mm). The use o a classic ixed ocus lens (endocen ic) is
no he mos sui able o measu emen asks like his. A
elecen ic lens and/o a lens using he Scheimp lug p inciple
o oblique plane sensing would be p e e able. Howe e ,
such wo iden ical lenses we e no a ailable o es ing. I is
also possible o selec mo e sui able came as wi h la ge chip
and esolu ion o eal applica ion.
The angle be ween he op ical axis o he i s came a and he
lase plane was app oxima ely 30° (α
1
= 30°) and o he
second came a α
2
= 55°. These angles we e chosen as a
comp omise be ween he desi ed ange and esolu ion in
Z-axis, he quali y o he e lec ed ligh s ip in he cap u ed
images, and a su icien di e ence be ween he wo angles
o success ul compensa ion o he oil laye . The measu ing
sys em o he assembled wo kplace had a pixel esolu ion in
he X-axis o app oxima ely 20 µm and in he Z-axis o he
i s came a 40 µm and o he second 24 µm.
The con e sion ela ions be ween he image coo dina es and
he eal coo dina e sys em in oduced o he measu ed objec
can be de e mined by calcula ion, bu i is necessa y o know
p ecisely all he equi ed inpu pa ame e s (e.g. pa ame e s o
came as and lenses, o ien a ion o came as, e c.) and in case
o some changes du ing use o measu ing sys em i is
necessa y o measu e all a iables again and ecalcula e he
ela ionships. O i is possible o de e mine he
in e ela ionships o indi idual iangula ion sys ems
(came a, lase , and measu ed objec ) by calib a ion. I a
sui able au oma ic p ocedu e and solu ion o calib a ion a e
designed, his app oach is mo e app op ia e because i allows
a con inuous ecalib a ion o he en i e measu ing sys em a
e e y change o e en p e en i ely (a sligh change in he
ela ionship can also occu due o en i onmen al in luences,
e.g. empe a u e). Calib a ion o expe imen al sys ems was
512 I. Janako a e al. / IFAC Pape sOnLine 52-27 (2019) 508–513
pe o med on se e al ings whose eal spa ial coo dina es
we e sub ac ed om p e iously measu ed p o iles by
ScanCon ol 2900-25/BL scanne . Mo e app op ia e, e en
wi h espec o he planned ecalib a ion, would be o use a
speci ically designed o his pu pose (sui able dimensions
and dis ibu ion o calib a ion poin s) and p ecisely
manu ac u ed calib a ion objec ha would bes co e he
en i e equi ed space wi h he calib a ion poin s. Fo he se
o poin s wi h known spa ial coo dina es (x, z) a e assigned
co esponding coo dina es (u, ) in he images. Two (one o
each came a) ans o ma ion ma ices (size 3x3) de e mined
by he leas squa es me hod a e he esul o he calib a ion.
Simila ly, he con e sion ela ionship be ween he oil
hickness and he e o p oduced can be de e mined by
calcula ion o calib a ion. Fo he same eason, howe e ,
hese ela ionships we e again de e mined by calib a ion.
Also because he e ac i e index o he pa icula used oil
was no exac ly known. Fo his pu pose, he oil- ee ings
we e i s measu ed and hen he oil o unknown hickness
was applied o hem and he di e ences in alues we e
de e mined by e-measu emen . Based on se e al
measu emen s, he cons an
K
ep esen ing he a e age e o
a io o he second and he i s came a
K
=
e
2
/
e
1
was
de e mined (a he cos o some inaccu acy in endocen ic
lenses). The ad an age o his app oach o calib a ion is ha
he e is no need o know he cu en oil hickness.
Fo he pu pose o measu ing bea ing p o iles, undis o ed as
well as dis o ed by oil laye , a p og am o sub-pixel lase
line de ec ion and da a il e ing had o be c ea ed. An
example o lase ack e alua ion is shown in Fig. 6. As can
be seen om he p esen ed igu e, he quali y o lase ace
acing is highly dependen on he su ace o he objec being
measu ed. P oblems a e caused by une en su aces, gla e,
less e lec i e and inclined su aced in he aceway, and
shadows be ween he bea ing and base (in he case o
elecen ic lens use, blind spo s and shadows should no be
c ea ed). A e applying he oil, he lase ack, especially in
he aceway, whe e he la ges amoun o oil ends o be, is
less no iceable and mo e blu ed. The quali y o lase
acking can be imp o ed by aking images wi h se e al
di e en exposu e imes and selec ing he mos app op ia e
image o each bea ing egion.
Fig. 6. The esul o he lase ack de ec ion p og am
A e lase ack de ec ion, he pixel posi ions o he line a e
ans o med a spa ial coo dina es based on calib a ion da a
o each came a. The ollowing igu e shows wo measu ed
p o iles o he same ing wi hou an oil laye om he i s
and he second came a. Ideally (especially pe ec line
de ec ion and calib a ion), bo h p o iles should be he same
a e ans o ma ion. Howe e , due o e.g. e lec ions, limi ed
and unequal came a esolu ion, e c., he eal ob ained p o iles
a e sligh ly di e en . The di e ences a e shown by he g een
do s in Fig. 7. In his pa icula case, he a e age p o iles
di e ence is 6.5 µm, he s anda d de ia ion is 12.7 µm.
Fig. 7. P o ile measu ed by i s came a (blue) and second
came a ( ed) and di e ence o alues (g een)
A e applying he oil, as expec ed, he measu ed da a is
de o med by he oil laye (ligh e ac ion). The measu ed
p o iles o he same oil- ee (blue) and oiled ( ed) bea ing a e
shown in he ollowing igu e. As you can see om he
di e ence o alues (g een), he e o a he deepes poin o
he aceway is o e 200 µm o he i s came a and nea ly
300 µm o he second (due o he la ge came a-lase angle).
This co esponds o he oil hickness o app oxima ely
0.55 mm. Howe e , i should be no ed ha his oil hickness
is un ealis ically high and se es only o p esen a ion. The
usual oil hickness a e p ese a ion and d ipping in
ope a ion is up o abou 100 µm.
Fig. 8. Measu ed p o ile o he same bea ing wi hou oil
(blue), a e oil applica ion ( ed) and di e ence (g een)
When wo oil-coa ed ing p o iles a e measu ed wi h came as
a di e en angles, he eal bea ing p o ile can be calcula ed
based also on he de o med da a hanks o he con e sion
ela ionships speci ied by calib a ion ( o each X-coo dina e,
he Z-coo dina e can be de e mined wha would be wi hou
oil). The calcula ion can be done using he o mula:
1
21
−
−
⋅
=
K
zKz
z
ee
, (4)

I. Janako a e al. / IFAC Pape sOnLine 52-27 (2019) 508–513 513
The esul i g aphed in Fig. 9, whe ein he calcula ed oil
laye compensa ion p o ile is shown in black. Fo he sake o
cla i y, he g aph is supplemen ed by p o iles measu ed by
bo h came as wi hou he oil laye (blue and ed wa e o ms
as in Fig. 7) and by he di e ence be ween he compensa ed
p o ile and he oil- ee p o ile om he i s came a (g een
wa e o m). Al hough he dispe sion o he di e ence
be ween he p o ile o he i s oil- ee came a and he
compensa ed p o ile is la ge in his case, as shown by he
la ge s anda d de ia ion alue o 52.8 µm, howe e , he
a e age di e ence is only -2.7 µm.
Fig. 9. Compensa ed p o ile (black) and di e ence (g een)
om he p o ile wi hou oil om i s came a (blue)
The oil laye , i no compensa ed, has a undamen al
in luence on he de e mina ion o he basic bea ing
pa ame e s - ing heigh and aceway adius, espec i ely
aceway dep h. Sepa a e algo i hms we e p og ammed o
de e mine hese pa ame e s. Fo example, he ing heigh is
de e mined by app oxima ion o da a a he ing's op edges
and a he basepla e by he s aigh lines. The da a in he
aceway, on he o he hand, is in e laced wi h a ci cle.
I he bea ing dimensions we e measu ed wi h only one
iangula ion sys em, wi hou he possibili y o elimina ing
he e ec o he laye , he measu ed bea ing heigh would be
highe han he eali y and he aceway adius would also be
la ge . Fig. 10 shows a de ail o p o ile wi h oil wi hou ( ed)
and a e (black) compensa ion. The calcula ed aceway adii
de e mined by i ing he aceway poin s wi h he ci cle a e
indica ed by he same colou . The blue pa o he ci cle
indica es he adius speci ied o he oil- ee bea ing. I can
be seen ha while he compensa ed p o ile has a g ea e
dispe sion o alues, he alues be e ollow he eal shape
Fig. 10. Raceway adius de e mina ion - p o ile wi hou ( ed)
and a e (black) compensa ion
o he aceway. Howe e , he sho sec ion o he ci cle is no
su icien o accu a ely de e mine he diame e o he aceway
bu will be used o de e mine i s dep h and o he bea ing
pa ame e s.
6. CONCLUSIONS
The a icle deals wi h he undamen al limi a ion o he
applica ion o he basic simple p inciple o he ac i e
iangula ion me hod o objec s ha a e co e ed by some
anspa en laye (e.g. p ese ing oil, lub ican , wa e ). Fo
his eason, he inno a i e sys em o measu e 3D pa ame e s
o objec s h ough his laye has been designed. The basic
idea is o use wo coope a ing senso s each wi h he di e en
angle o he lase . The ad an age is ha he same sys em can
be used o ano he ype o ask wi h wo unknowns (e.g. he
laye hickness o unknown e ac i e index o ice e sa).
Al hough he mos app op ia e ha dwa e was no a ailable
o he expe imen s, and e en he so wa e could no ye be
esol ed a he le el o he de elopmen o a comme cial
inished de ice, ne e heless he pe o med es s demons a e
ha he p oposed p inciple is co ec and ha his p ocedu e
is applicable o elimina e he anspa en laye .
ACKNOWLEDGEMENT
The comple ion o his pape was made possible by he g an
TH02010830 – “Au oma ion, measu emen and moni o ing
o assembly p ocesses” which is unded by Technology
Agency o he Czech Republic - p og amme EPSILON and
by he g an No. FEKT-S-17-4234 - „Indus y 4.0 in
au oma ion and cybe ne ics” inancially suppo ed by he
In e nal science und o B no Uni e si y o Technology.
REFERENCES
Amann, M.-Ch., Bosch, T., Lescu e, M., Myllylä, R. and
Rioux, M. (2001). Lase Ranging: A C i ical Re iew o
Unusual Techniques o Dis ance Measu emen . Op ical
Enginee ing, Vol. 40(1), pp. 10-19.
Be aldin, J.-A., Blais, F., Cou noye , L., Godin, G., Rioux,
M. and Taylo , J. (2003). Ac i e 3D Sensing. In: The e-
Way in o he Fou Dimensions o Cul u al He i age
Cong ess. Vienna, Aus ia.
Be ko ic, G. and Sha i , E. (2012). Op ical me hods o
dis ance and displacemen measu emen s. Ad ances in
Op ics and Pho onics, Vol. 4, Issue 4, pp. 441-471.
Geng, J. (2011). S uc u ed-ligh 3D su ace imaging:
a u o ial. Ad ances in Op ics and Pho onics, Vol. 3,
Issue 2, pp. 128-160.
Kalo á, I. and Lisz wan, M. (2005). Ac i e T iangula ion
Technique. In: 5 h In e na ional Con e ence o PhD
S uden s, pp. 99-104. Uni e si y o Miskolc, Miskolc.
Kalo á, I. and Lisz wan, M. (2006). Indus ial Applica ions
o T iangula ion Technique. IFAC P oceedings Volumes,
Vol. 39(21), pp. 258-263.
Schwa e, R., Heinol, H., Buxbaum, B., Ringbeck, T., Xu, Z.
and Ha mann, K. (1999). P inciples o Th ee-
Dimensional Imaging Techniques. In: Handbook o
Compu e Vision and Applica ions, Vol. 1, pp. 463-484.
Academic P ess, San Diego.