Тестування бюджетних фотосистем з метою досягнення оптимального результату для цілей оцифрування гербарію
The current paper stresses the application of different budget photosystems for digitization of herbarium specimens. Twelve photosystems were compared by color accuracy reproduction of the images. It was found that the photosystem built on the basis of photocamera Canon EOS 800D and fixed lens Tokin...
Збережено в:
| Дата: | 2023 |
|---|---|
| Автори: | , , , |
| Формат: | Стаття |
| Мова: | Англійська |
| Опубліковано: |
M.M. Gryshko National Botanical Garden of the NAS of Ukraine
2023
|
| Онлайн доступ: | https://www.plantintroduction.org/index.php/pi/article/view/1631 |
| Теги: |
Додати тег
Немає тегів, Будьте першим, хто поставить тег для цього запису!
|
| Назва журналу: | Plant Introduction |
| Завантажити файл: |  |
Репозитарії
Plant Introduction| _version_ | 1860145147621146624 |
|---|---|
| author | Novikov, Andriy Sup-Novikova, Mariia Nachychko, Viktor Kuzyarin, Oleksandr |
| author_facet | Novikov, Andriy Sup-Novikova, Mariia Nachychko, Viktor Kuzyarin, Oleksandr |
| author_sort | Novikov, Andriy |
| baseUrl_str | https://www.plantintroduction.org/index.php/pi/oai |
| collection | OJS |
| datestamp_date | 2024-04-07T19:57:15Z |
| description | The current paper stresses the application of different budget photosystems for digitization of herbarium specimens. Twelve photosystems were compared by color accuracy reproduction of the images. It was found that the photosystem built on the basis of photocamera Canon EOS 800D and fixed lens Tokina AT-X M35 PRO DX AF 35 mm f/2.8 Macro and currently used for the digitization of the LWS herbarium collection demonstrated the best results among other tested photosystems. It also produced the images with the same or even better color accuracy as in images downloaded from P, PI, B, and W virtual herbaria. Nevertheless, its color accuracy measured as ΔE2000, in general, does not meet recent criteria ascertained for the herbarium digitization purposes and new FADGI’s requirements. This photosystem has been found to have improving potential and, hence, should be optimized. On the other hand, it was also shown that smartphone Samsung Galaxy S10 could produce images with the same or even better color accuracy compared to some images deposited at P, PI, W, and B virtual herbaria. Therefore, in conditions of extreme situations and hostilities, such smartphones with additional external illumination can serve for urgent digitization of natural history collections. Finally, we doubt the application of commonly used color targets for the digitization of herbariaum preserved specimens since the original living color of such specimens is usually lost during conservation and preservation. Instead, it would be more beneficial to apply advanced targets to evaluate the spatial accuracy of images since they can incorrectly represent the important morphological characters of preserved specimens. |
| doi_str_mv | 10.46341/PI2023010 |
| first_indexed | 2025-07-17T12:54:18Z |
| format | Article |
| fulltext |
© The Authors. This content is provided under CC BY 4.0 license.
Plant Introduction, 99/100, 36–50 (2023)
RESEARCH ARTICLE
Testing budget photosystems to reach an optimal solution for the
herbarium digitization purposes
Andriy Novikov 1, *, Mariia Sup-Novikova 2, Viktor Nachychko 3, Oleksandr Kuzyarin 1
1 State Museum of Natural History, National Academy of Sciences of Ukraine, Teatralna str. 18, 79008 Lviv, Ukraine;
* novikoffav@gmail.com
2 Ukrainian Catholic University, Stryiska str. 29A, 79026 Lviv, Ukraine
3 Ivan Franko National University of Lviv, Hrushevskoho str. 4, 79005 Lviv, Ukraine
Received: 03.11.2023 | Accepted: 20.11.2023 | Published online: 24.11.2023
Abstract
The current paper stresses the application of different budget photosystems for digitization of herbarium
specimens. Twelve photosystems were compared by color accuracy reproduction of the images. It was
found that the photosystem built on the basis of photocamera Canon EOS 800D and fixed lens Tokina
AT-X M35 PRO DX AF 35 mm f/2.8 Macro and currently used for the digitization of the LWS herbarium
collection demonstrated the best results among other tested photosystems. It also produced the images
with the same or even better color accuracy as in images downloaded from P, PI, B, and W virtual herbaria.
Nevertheless, its color accuracy measured as ΔE2000, in general, does not meet recent criteria ascertained
for the herbarium digitization purposes and new FADGI’s requirements. This photosystem has been found
to have improving potential and, hence, should be optimized. On the other hand, it was also shown that
smartphone Samsung Galaxy S10 could produce images with the same or even better color accuracy
compared to some images deposited at P, PI, W, and B virtual herbaria. Therefore, in conditions of extreme
situations and hostilities, such smartphones with additional external illumination can serve for urgent
digitization of natural history collections. Finally, we doubt the application of commonly used color targets
for the digitization of herbariaum preserved specimens since the original living color of such specimens is
usually lost during conservation and preservation. Instead, it would be more beneficial to apply advanced
targets to evaluate the spatial accuracy of images since they can incorrectly represent the important
morphological characters of preserved specimens.
Keywords: herbarium digitization, image quality assessment, photosystems, delta E, color accuracy
https://doi.org/10.46341/PI2023010
UDC 57.082 + 57.087.3
Authors’ contributions: Andriy Novikov: conceptualization, project administration, supervision, funding acquisition, validation,
visualization, writing – original draft. Mariia Sup-Novikova: conceptualization, investigation, writing – original draft. Viktor Nachychko:
writing – original draft. Oleksandr Kuzyarin: writing – review & editing.
Funding: This work has been realized in the frames of the project “Digitization of natural history collections damaged as a result of
hostilities and related factors: development of protocols and implementation on the basis of the State Museum of Natural History
of the National Academy of Sciences of Ukraine” (Nr 2022.01/0013), financed by the National Research Foundation of Ukraine in the
grant call “Science for the Recovery of Ukraine in the War and Post-War Periods”.
Competing Interests: The authors declare that they have no conflict of interest.
https://creativecommons.org/licenses/by/4.0/
https://orcid.org/0000-0002-0112-5070
https://orcid.org/0000-0002-8542-3605
https://orcid.org/0000-0001-6756-2823
https://orcid.org/0000-0002-7728-3665
Plant Introduction • 99/100 37
Budget photosystems for the herbarium digitization purposes
Introduction
The digitization of natural history collections
has attracted the attention of many recent
specialists (Nelson & Ellis, 2019; Hedrick
et al., 2020; Hussein et al., 2022). At its best, it
grants free, fast, and easy access to deposited
specimens through databases and virtual
collections (Drew et al., 2017; Schindel & Cook,
2018; Jackowiak et al., 2022). Such virtual
collections do not fully substitute original
deposited material but significantly extend its
application in research (Gries et al., 2014; Khan
et al., 2021; Davis, 2023), including extensive
biodiversity-related meta-analyses (Dawson-
Glass & Hargreaves, 2022; Carta et al., 2022;
Samain et al., 2023) and, in some sense,
secure its long-term virtual preservation
(Novikov, 2019). Such virtual backup of the
natural history collections in Ukraine became
a crucial task in light of the negative effect of
hostilities and related factors, e.g., the inability
to warrant stable preservation conditions
due to blackouts or lack of personnel (Vajda
et al., 2022; Novikov et al., 2023). Being at risk
of direct attack or air strike, every natural
history collection in Ukraine is insecure and
can be damaged (Mosyakin & Shiyan, 2022) or
destroyed (Pavlyshyn, 2022; Le Page, 2022).
Image quality assessment is one of the
steps in the digitization workflow (Nieva de la
Hidalga et al., 2020). The quality of obtained
digital images can be evaluated based on
numerous parameters, including colorimetric
ones. Among them, color reproduction quality
is an important parameter that is considered
during the digitization of natural history
collections (Hasler & Suesstrunk, 2003;
Zerman et al., 2019). It is usually quantified
as ΔE (delta E) – the difference between
the actual color of the object and the color
reproduced on the image (Hasler & Suesstrunk,
2003; Palus, 2006; Gu et al., 2022). The
smallest ΔE conforms to the best reproduction
of the color. The acceptable ΔE value is
regulated by two leading legal authorities, i.e.,
Metamorfoze (the Dutch National Programme
for the Preservation of Paper Heritage) and
FADGI (the Federal Agencies Digitization
Guidelines Initiative, USA). Among other image
parameters (colorimetric and representing
spatial accuracy), Metamorfoze (van Dormolen,
2012, 2019) calculates ΔE using the technical
test chart Digital ColorChecker SG with an old
CIE 1976 formulae (ΔE1976). In the recent FADGI
guide (Rieger et al., 2023), the calculation of
ΔE follows ISO 13658 : 2000 standard (ΔE2000).
The application of different formulae makes
these two standards incomparable directly
but allows ascertaining some parallels.
Metamorfoze has three levels of image
quality (from low to high) based on complex
calculations, i.e., Extra Light, Light, and Strict.
Similarly to Metamorfoze, FADGI delimits four
grades of digital image quality. The lowest
FADGI’s one-star level has no equivalent level
in Metamorfoze, while the next two- to four-
star levels more or less correspond to Extra
Light, Light, and Strict Metamorfoze levels,
respectively (DT Heritage, 2023).
Neither Metamorfoze nor FADGI regulate
the digitization of natural history collections
directly but provide general requirements
for digital images of the heritage. FADGI
(Rieger, 2016; Rieger et al., 2023) subdivides
the digitized materials into several types, and
each type has its own four levels of quality
with ascertained requirements, including
acceptable ΔE color accuracy level. Following
the recently published FADGI guide (Rieger
et al., 2023), the digitization of herbarium
collections, for example, falls under the
definition “Documents (Unbound): General
Collections”. Meanwhile, in Metamorfoze (van
Dormolen, 2012), there is no initial subdivision
of digitized materials by type, and only three
levels of quality are ascertained with the
indication of materials for digitization of
which they are applied. Hence, the digitization
of herbarium collections best fits the
Metamorfoze Light level.
The Metamorfoze, in a strict sense (van
Dormolen, 2012), requires the average ΔE1976
for digital images to be ≤4 while the maximum
ΔE1976 should not exceed 10 (van Dormolen,
2012). However, there is also a Metamorfoze
Light and Metamorfoze Extra Light levels,
which require average ΔE1976 ≤ 5 and maximum
ΔE1976 ≤ 18. In FADGI (Rieger et al., 2023), the
lowest ΔE2000 (one-star quality level) should
not exceed 6.5, and the maximum ΔE should
not exceed 13. For FADGI’s highest quality,
four-star level, the average ΔE2000 should not
exceed 2, and the maximum ΔE2000 should not
exceed 4. Considering FADGI’s standard, Nieva
de la Hidalga et al. (2020) suggested that for the
herbarium material, the minimal acceptable
ΔE2000 value of images should be < 5.
38 Plant Introduction • 99/100
Novikov et al.
This study aimed to stress the color quality
of images obtained using the photosystem
built combining cropped photocamera
Canon EOS 800D and fixed lens Tokina AT-X
M35 PRO DX AF 35 mm f/2.8 Macro. This
photosystem is currently used for digitization
of the herbarium of vascular plants at the
State Museum of Natural History of the NAS
of Ukraine (LWS) and is described in detail by
Novikov & Sup-Novikova (2021).
Material and methods
We tested 12 camera/lens combinations
of different classes (Table 1) placed on the
horizontal tripod Beike Q999H with two
LED lamps Yongnuo YN-300 Air at the same
illumination conditions. For all cameras,
the same or closest possible presets were
applied, including light sensitivity ISO 200,
diaphragm f/5.6, and exposition 1/50 s. The
ColorChecker Classic Mini, regularly used
for image capturing during the herbarium
digitization, has been applied as a test target.
Obtained images in jpeg format were not
objected for any postprocessing and analyzed
as is. For this purpose, their ΔE1976 and ΔE2000
values were calculated using delt.ae free web
application (Picturae, 2023).
We also compared the ΔE2000 of the images
captured by the LWS’s regular photosystem
(i.e., Canon EOS 800D with Tokina AT-X M35
PRO DX AF 35 mm f/2.8 Macro) with those
for images obtained from ten virtual herbaria,
namely BR (jpeg), L ( jpeg), P ( jpeg), B ( jpeg),
US (jpeg), NY (jpeg), W (tiff), BM (tiff and
jpeg), GZU (tiff), and PI (tiff). The acronyms
of represented herbaria follow Thiers (2023).
In particular, 30 randomly selected images
were downloaded from each mentioned digital
herbaria in the highest available resolution
and processed through the delt.ae free web
application (Picturae, 2023). The images from
BM herbarium were downloaded both in jpeg
and tiff formats. Thus, including LWS, 360
images were analyzed in total. After that,
outliers were excluded from the analyses
and substituted by additionally downloaded
images. Minimum, maximum, and mean
values, as well as standard deviation (SD) for
obtained ΔE2000 were calculated, and graphs
were visualized using PAST ver. 4.14 software
(Hammer et al., 2001).
Results and discussion
Among all tested camera/lens
combinations, the photosystem currently
applied at the LWS herbarium demonstrated
the best result and was the only producing
part of images meeting Nieva de la Hidalga
et al. (2020) criteria. Following the recently
published FADGI guide (Rieger et al., 2023),
LWS’s photosystem received only two stars
among four. However, after the exclusion
of outliers, the LWS’s photosystem
demonstrated only 7.7 mean value of ΔE2000,
and hence did not pass FADGI (Rieger et al.,
2023) and Nieva de la Hidalga et al. (2020)
requirements (Table 1).
Besides this, only the full-frame Sony
α7 II with lens FE 28-70 mm 3.5-5.6 OSS
passed the new FADGI requirements (Rieger
et al., 2023) and received one star without
processing images through CameraRAW
app. Not all other tested photosystems
passed the FADGI requirements on minimal
acceptable value of ΔE2000. Following the
currently applied Metamorfoze guide
(van Dormolen, 2012), none of the tested
photosystems meet minimal ΔE1976 value
requirements (Table 1).
Interestingly, the inbuilt camera of
the old smartphone Samsung Galaxy S10
camera demonstrated better color accuracy
than some cameras with stock or integrated
lenses (Table 1). However, it should be
taken into account that images produced
by smartphones are processed through
the original firmware and different apps.
As a result, the quality of obtained images
can differ even between the same models
of smartphones using different versions of
the firmware or different apps. To avoid
serious modification of the images by the
original firmware of the smartphone, it is
recommended to use the Open Camera app.
It is also interesting that Canon EOS
800D with a cropped matrix but with a high-
end macro lens from Tokina demonstrated
better color accuracy than the full-frame
camera Sony α7 II with a stock lens. Hence,
the crop factor is not limiting here, and
the primary attention in the photosystem
construction for digitization purposes
should be paid to the lens quality.
The same cameras with different lenses
demonstrated quite different ΔE outcomes
Plant Introduction • 99/100 39
Budget photosystems for the herbarium digitization purposes
Ta
bl
e
1.
T
he
b
es
t v
al
ue
s
of
∆
E 20
00
o
bt
ai
ne
d
us
in
g
di
ff
er
en
t t
es
tin
g
ph
ot
os
ys
te
m
s.
C
am
er
a
an
d
le
ns
s
pe
cs
C
an
on
P
ow
er
S
ho
t
SX
10
IS
w
ith
IS
5
-1
00
m
m
1:
2.
8-
5.
7
U
SM
N
ik
on
1
V1
w
ith
N
ik
ko
r
10
-3
0
m
m
1:3
.5
-5
.6
V
R
C
an
on
R
eb
el
T
7i
(E
O
S
80
0D
) w
ith
C
an
on
E
F
50
m
m
f/
1.4
U
SM
C
an
on
R
eb
el
T
4i
(E
O
S
65
0D
) w
ith
EF
S
18
-5
5
m
m
3
.5
-
5.
6
IS
S
TM
C
an
on
R
eb
el
T
4i
(E
O
S
65
0D
) w
ith
EF
S
18
-1
35
m
m
1:3
.5
-5
.6
IS
U
SM
Sa
m
su
ng
E
X2
F
w
ith
S
ch
ne
id
er
-
Kr
eu
zn
ac
h
Va
ri
op
la
n
Zo
om
5.
2-
17
.2
m
m
1:
1.4
-2
.7
D
ua
ls
C
am
er
a
cl
as
s
O
ld
te
le
-z
oo
m
ca
m
er
a
O
ld
m
ir
ro
rl
es
s
ca
m
er
a
Re
ce
nt
c
ro
pp
ed
ca
m
er
a
O
ld
c
ro
pp
ed
c
am
er
a
O
ld
c
ro
pp
ed
c
am
er
a
O
ld
e
xt
ra
-b
ri
gh
t
po
ck
et
c
am
er
a
C
ol
or
re
pr
od
uc
tio
n
ac
cu
ra
cy
fr
om
lo
w
to
h
ig
h
Se
ns
or
ty
pe
C
C
D
C
M
O
S
C
M
O
S
C
M
O
S
C
M
O
S
BS
I C
M
O
S
Se
ns
or
s
iz
e
1/
2.
3”
(6
.17
×
4
.5
5
m
m
)
1″
(1
3.
2
×
8.
8
m
m
)
AP
S-
C
(2
2.3
×
14
.9
m
m
)
AP
S-
C
(2
2.3
×
14
.9
m
m
)
AP
S-
C
(2
2.3
×
14
.9
m
m
)
1/
1.7
" (
7.6
×
5
.7
m
m
)
Re
so
lu
tio
n
10
M
p
10
M
p
24
.2
18
M
p
18
M
p
12
.4
M
p
Be
st
µ
∆
E 20
00
14
.3
10
.8
10
.4
9.
7
9.
7
9.
2
Be
st
µ
∆
E 19
76
18
.3
22
.5
13
.9
20
.8
13
.8
12
.4
D
el
t.a
e
C
ol
or
E
nc
od
in
g
Er
ro
r
(∆
E 20
00
)
*
*
*
*
*
*
FA
D
G
I E
d.
2
(R
ie
ge
r,
20
16
)
–
–
*
*
*
*
FA
D
G
I E
d.
3
(R
ie
ge
r
et
a
l.,
2
02
3)
–
–
–
–
–
–
M
et
am
or
fo
ze
(v
an
D
or
m
ol
en
, 2
01
2)
–
–
–
–
–
–
Ac
ce
pt
ab
le
le
ve
l f
or
th
e
he
rb
ar
iu
m
di
gi
tiz
at
io
n
(N
ie
va
d
e
la
H
id
al
ga
et
a
l.,
2
02
0)
N
o
N
o
N
o
N
o
N
o
N
o
40 Plant Introduction • 99/100
Novikov et al.
C
am
er
a
an
d
le
ns
s
pe
cs
Sa
m
su
ng
G
al
ax
y
S1
0
Pa
na
so
ni
c
Lu
m
ix
D
C-
G
9
w
ith
Si
gm
a
30
m
m
f/
1.4
D
C
D
N
C
on
te
m
po
ra
ry
(r
eg
ul
ar
m
od
e)
N
ik
on
D
90
w
ith
N
ik
ko
r
D
X
AF
-S
18
-1
05
m
m
1:
3.
5-
5.
6
G
E
D
C
an
on
R
eb
el
T7
i (
EO
S
80
0D
)
w
ith
C
an
on
E
F
10
0
m
m
f/
2,
8L
M
ac
ro
IS
U
SM
So
ny
A
lp
ha
IL
C
E-
α7
II
w
ith
FE
2
8-
70
m
m
3.
5-
5.
6
O
SS
(IL
C
E-
7M
2)
Pa
na
so
ni
c
Lu
m
ix
D
C-
G
9
w
ith
Si
gm
a
30
m
m
f/
1.4
D
C
D
N
C
on
te
m
po
ra
ry
(p
ix
el
-s
hi
ft
m
od
e)
C
an
on
R
eb
el
T
7i
(E
O
S
80
0D
) w
ith
To
ki
na
M
ac
ro
3
5
F2
.8
m
m
D
X
AT
-X
Pr
o
C
am
er
a
cl
as
s
Sm
ar
tp
ho
ne
Re
ce
nt
m
ir
ro
rl
es
s
ca
m
er
a
O
ld
c
ro
pp
ed
ca
m
er
a
Re
ce
nt
c
ro
pp
ed
ca
m
er
a
Fu
ll-
fr
am
e
ca
m
er
a
Re
ce
nt
m
ir
ro
rl
es
s
ca
m
er
a
Re
ce
nt
c
ro
pp
ed
ca
m
er
a
C
ol
or
re
pr
od
uc
tio
n
ac
cu
ra
cy
fr
om
lo
w
to
h
ig
h
Se
ns
or
ty
pe
IS
O
C
EL
L
C
M
O
S
C
M
O
S
C
M
O
S
C
M
O
S
C
M
O
S
C
M
O
S
C
M
O
S
Se
ns
or
s
iz
e
SA
K2
L4
1/
2.
55
"
(5
.6
×
4
.2
m
m
)
M
ic
ro
4
/3
(1
7.3
×
13
.0
m
m
)
AP
S-
C
(2
2.
3
×
14
.9
m
m
)
AP
S-
C
(2
2.
3
×
14
.9
m
m
)
Fu
ll-
Fr
am
e
35
m
m
(3
5.8
×
23
.9
m
m
)
M
ic
ro
4
/3
(1
7.3
×
13
.0
m
m
)
AP
S-
C
(2
2.
3
×
14
.9
m
m
)
Re
so
lu
tio
n
12
M
p
20
.3
M
p
12
.3
M
p
24
.2
24
.3
M
p
80
.6
M
p
24
.2
Be
st
µ
∆
E 20
00
8.
1
7.7
7.6
7.5
6
5.
7
4.
7
Be
st
µ
∆
E 19
76
11
.3
15
.2
9.
7
10
.1
13
.9
12
.8
7.8
D
el
t.a
e
C
ol
or
E
nc
od
in
g
Er
ro
r
(∆
E 20
00
)
**
**
**
**
**
**
**
*
FA
D
G
I E
d.
2
(R
ie
ge
r,
20
16
)
*
**
**
**
**
**
**
*
FA
D
G
I E
d.
3
(R
ie
ge
r
et
a
l.,
2
02
3)
–
–
–
–
*
**
**
M
et
am
or
fo
ze
(v
an
D
or
m
ol
en
, 2
01
2)
–
–
–
–
–
–
–
Ac
ce
pt
ab
le
le
ve
l f
or
th
e
he
rb
ar
iu
m
di
gi
tiz
at
io
n
(N
ie
va
d
e
la
H
id
al
ga
et
a
l.,
2
02
0)
N
o
N
o
N
o
N
o
N
o
Ye
s
af
te
r
po
st
pr
oc
es
si
ng
Ye
s
Ta
bl
e
1.
C
on
tin
ue
d.
Plant Introduction • 99/100 41
Budget photosystems for the herbarium digitization purposes
(Table 1). In general, cameras with macro
lenses, usually having smaller aperture and
better sharpness, also demonstrated better
color accuracy. Interestingly, the recent
mirrorless camera Lumix G9 in the testing
set (ISO 200, f/5.6, 1/50 s) showed quite
a low color accuracy in regular mode, and
only after the application of pixel shifting
(high-resolution mode) did it appear nearly
close to the best results demonstrated by
Canon 800D with Tokina Macro 35 mm lens.
However, after application of other settings
(ISO 200, f/8, 1/15 s, ISO 200, f/5.6, 1/30 s,
and ISO 400, f/5.6, 1/60 s) it stable produced
images with ΔE2000 values 4.6, 4.9, and 5.0,
respectively. After the basic postprocessing
in the CameraRAW application, it showed
even better results with ΔE2000 values 4.4, 4.6,
and 4.7, respectively. Considering also the
higher final resolution of the images (80.6
Mp), Lumix G9 with Olympus 30 mm macro
lens seems to be even more perspective
photosystem, which will be tested deeply
in our further research. At the moment,
we can assume that precise choice of the
lens and test of different settings in the
diapason of ISO 200–400, diaphragm f/4–8,
and exposition 1/15–1/60 s is an essential
step in setting up the photosystem.
Some authors (e.g., iDigBio. 2014; Harris
& Marsico, 2017; Takano et al., 2019)
recommend applying ISO 100 and smaller
exposition (up to 1/200 s). We found that
this highly depends on the illumination,
and even ISO 400 results in good-quality
images. Similarly, ISO 400 is proposed for
digitization purposes by Baratè et al. (2020).
As we found, it is better to avoid aperture
values higher than f/3.5 as they result in
images with somewhat blurred contours
of specimens and more vignetting. It is
also good to avoid the apertures below f/8
since they often result in arising chromatic
aberrations. ISO over 400 should be avoided
since it results in losing image details and
increased noise.
To stress the requirements of Nieva
de la Hidalga et al. (2020), who declared a
minimally acceptable level of ΔE2000 < 5 for
all kinds of image use, including Internet
publishing and preservation, we downloaded
random images from ten virtual herbaria
and compared calculated ΔE2000 values with
those obtained using LWS’s photosystem.
The raw obtained data are represented in
Appendix A. The basic statistics on obtained
data are represented in Table 2, where the
herbaria consulted are organized from the
lowest to highest demonstrated ΔE2000 value.
The same statistics have been visualized on
the graph (Fig. 1). As we can see, only two
herbaria (L and NY) provide images with
ΔE2000 values lower than 5, hence, meeting
the criteria of Nieva de la Hidalga et al.
(2020). US, GZU, BM, BR, and P provide only
some images having ΔE2000 < 5, while other
images demonstrate significantly higher
ΔE2000 values. The remaining herbaria, W,
B, PI, and LWS (with minor exceptions),
provide images with higher ΔE2000 values.
Interestingly, the images from P, PI,
W, and B virtual herbaria show the mean
ΔE2000 value, which is similar or even higher
compared to that in the photos obtained
by the camera built into the smartphone
Samsung Galaxy S10. This does not mean
that smartphones can be generally accepted
for digitization purposes. It instead points
to the progress in the quality of digital
cameras, which is also reflected in the
stricter requirements of the current FADGI
guide (Rieger et al., 2023) compared to its
previous edition (Rieger, 2016). However,
we suggest that in extreme situations,
like in the case of hostilities, smartphones
with good-quality cameras can serve as
an alternative for the urgent digitization
of natural history collections. Such urgent
virtual back-upping of the collections
could be the only chance to save them.
Nevertheless, in such a case, attention
should be paid to the quality of illumination
that cannot be provided by smartphone LED
flash. Hence, external illuminators (at least
two oppositely placed household LED lamps
with 6500 K) are strictly required.
Hence, if taking into account the
complete set of tested images from different
virtual herbaria, most of them do not meet
the color accuracy criteria set by Nieva de
la Hidalga et al. (2020). The digital images of
herbarium specimens of any resolution and
quality cannot fully substitute the original
material for research purposes. They
serve rather for preliminary evaluation
and validation of the data provided on
specimens’ labels. Hence, an acceptable
ΔE2000 value, perhaps, should be extended to
42 Plant Introduction • 99/100
Novikov et al.
C
ha
ra
ct
er
is
tic
s
L
(jp
eg
)
N
Y
(jp
eg
)
U
S
(jp
eg
)
G
ZU
(t
iff
)
BM
(t
iff
)
BM
(j
pe
g)
BR
(j
pe
g)
LW
S
(jp
eg
)
P
(jp
eg
)
PI
(t
iff
)
W
(t
iff
)
B
(jp
eg
)
N
30
30
30
30
30
30
30
30
30
30
30
30
M
in
1.4
5
3.
56
3.
23
4.
27
4.
36
6.
10
2.
79
6.
44
3.
98
5.
97
5.
35
5.
73
M
ax
2.
75
5.
72
7.0
4
6.
53
7.0
2
9.
78
11
.5
5
9.
98
11
.9
6
11
.3
0
16
.3
9
15
.6
2
M
ea
n
1.9
4
4.
62
5.
25
5.
31
5.
67
8.
48
6.
86
7.6
5
8.
06
8.
55
9.
68
10
.9
1
St
an
da
rd
d
ev
ia
tio
n
0.
38
0.
61
1.0
9
0.
62
0.
74
1.2
3
2.
26
0.
95
2.
55
1.5
2
3.
52
2.
53
M
ed
ia
n
1.7
7
4.
42
5.
45
5.
21
5.
73
9.
00
7.3
3
7.4
3
8.
28
8.
55
9.
02
10
.9
4
C
oe
ffi
ci
en
t o
f
va
ri
at
io
n
19
.8
5
13
.2
6
20
.8
4
11
.5
9
13
.0
5
14
.5
4
33
.0
2
12
.3
9
31
.6
9
17
.77
36
.3
5
23
.2
3
C
ol
or
a
cc
ur
ac
y
le
ve
ls
of
im
ag
es
fo
llo
w
in
g
FA
D
G
I E
d.
3
(R
ie
ge
r
et
al
.,
20
23
)
**
**
, *
**
**
, *
**
*,
*
*,
*,
–
**
, *
**
, *
, –
*,
–
**
*,
*
*,
*,
–
**
(o
ut
lie
rs
),
*,
–
**
, *
, –
*,
–
*,
–
*,
–
M
ea
n
co
lo
r
ac
cu
ra
cy
le
ve
l o
f i
m
ag
es
fo
llo
w
in
g
FA
D
G
I E
d.
3
(R
ie
ge
r
et
a
l.,
2
02
3)
**
**
**
*
*
*
–
–
–
–
–
–
–
Ac
ce
pt
ab
le
le
ve
l
fo
r
th
e
he
rb
ar
iu
m
di
gi
tiz
at
io
n
(N
ie
va
d
e
la
H
id
al
ga
e
t a
l.,
2
02
0)
Ye
s
Ye
s
Pa
rt
ly
Pa
rt
ly
Pa
rt
ly
N
o
Pa
rt
ly
N
o
if
ou
tli
er
s
ex
cl
ud
ed
Pa
rt
ly
N
o
N
o
N
o
Ta
bl
e
2.
B
as
ic
s
ta
tis
tic
s
on
Δ
E 20
00
v
al
ue
fo
r i
m
ag
es
p
ro
vi
de
d
by
d
iff
er
en
t h
er
ba
ri
a
an
d
its
c
or
re
sp
on
de
nc
e
to
c
ur
re
nt
q
ua
lit
y
re
qu
ir
em
en
ts
. B
ol
d
fo
nt
in
di
ca
te
s
th
e
ΔE
20
00
<
5
.
Plant Introduction • 99/100 43
Budget photosystems for the herbarium digitization purposes
6.5, which corresponds to the FADGI’s one-
star quality level (Rieger et al., 2023).
The variation of quality of provided
images has been observed in all virtual
herbaria consulted but is the highest in W,
BR, and P. It is perhaps due to the long period
of digitization of their collections and/or
application of photosystems of updating
quality during this period. It is worth noting
that the coefficient of variation of ΔE2000 for
LWS’s images is one of the smallest among
tested herbaria, which points to the stable
quality of obtained images due to the use of
the same protocol and photosystem.
Only L herbarium provides images
corresponding to FADGI’s four-star and
three-star levels (Rieger et al., 2023)
with ΔE2000 < 2 (Table 2; Fig. 1). US and BR
herbaria provide some images of three-star
level, but in general have images of one-
star level or out of acceptable range (i.e.,
with ΔE2000 > 6.5). Herbarium NY provides
most images corresponding to FADGI’s
two-star level with ΔE2000 < 5, and some
images of one-star level with ΔE2000 < 6.5.
Similarly, GZU, BM, and P herbaria provide
some images of two-star level, but on
average, they have only one-star images or
are out of the acceptable range. PI, W, and
B herbaria provide some images of one-
star level, but generally, their images do
not meet FADGI’s criteria regarding color
accuracy. By analogy, LWS herbarium has
only exceptional images of two star-level,
while most images meet only one-star
level or do not meet FADGI’s criteria. In
conclusion, the color accuracy of images
provided by LWS is similar to those in PI, W,
and B herbaria and very close to those in P
and BR, but generally requires improvement
to meet recent standards.
The color accuracy, along with bit
depth, is one of the main criteria Nieva
de la Hidalga et al. (2020) ascertained for
digitizing herbarium materials. However,
it is unclear why much attention is paid to
color accuracy reproduction if preserved
specimens are already dried and lost their
Figure 1. The violin and box plot graph demonstrating the variation of ΔE2000 value for images provided by
different herbaria. Whiskers represent standard deviations; central lines at notches represent medians.
44 Plant Introduction • 99/100
Novikov et al.
living natural colors in herbarization and
during further preservation. Probably,
more attention should be paid to the
spatial accuracy of images of the preserved
specimens, including herbarium ones. The
resolution, sharpness (postprocessing
sharpening of master images is restricted
by Metamorfoze and FADGI – van
Dormolen, 2012; Rieger et al., 2023),
geometric distortion, and presence of
artifacts are more important criteria for
the digitization of preserved specimens,
since they can affect the morphological
characters important for taxonomic work
or identification of digitized specimens.
However, the spatial accuracy of obtained
images cannot be evaluated using regular
color targets (e.g., ColorChecker or Kodak
Color Control Patches) commonly applied
during digitization. To assess the spatial
accuracy of obtained images, advanced
targets like Golden Thread Object Level
Target (applied at the herbarium L) or
Image Engineering Scan Reference Chart
TE263 (applied at the herbarium KRAM) are
necessary.
Conclusions
The current LWS’s photosystem
demonstrated the best results regarding
color accuracy reproduction compared to
other tested photosystems. It was shown that
the full-frame camera does not guarantee
better color accuracy of obtained images,
and the primary attention should be paid
to the lens quality instead. Nevertheless,
the current LWS’s photosystem produces
most images with color accuracy that does
not meet recent FADGI’s (Rieger et al., 2023)
minimal requirements (ΔE2000 < 6.5) and
criteria ascertained by Nieva de la Hidalga
et al. (2020) for the herbarium digitization
purposes (ΔE2000 < 5). It can produce images
of better color accuracy (ΔE2000 = 4.7),
meeting the mentioned criteria, but this
requires further optimization and camera
adjustment. On the other hand, it produces
images of the same or even better quality
than those provided by some virtual
herbaria (i.e., P, PI, B, and W). Hence,
there also arises the question about the
acceptable limit of ΔE2000, which probably
should be leveled to 6.5 compared to 5 units
ascertained by Nieva de la Hidalga et al.
(2020).
It was also shown that the Samsung
Galaxy S10 smartphone’s camera could
produce images with the same or even
better color accuracy compared to some
images from P, PI, W, and B herbaria. Hence,
we suggest that in extreme situations, like
in the case of hostilities, smartphones
with good quality cameras and additional
external illumination can serve for
urgent digitization of the natural history
collections.
Finally, we doubt the application of
commonly used color targets for digitizing
preserved specimens since the original
coloration of such specimens is usually lost.
Instead, it would be more useful to apply
targets for evaluating the spatial accuracy
of obtained images since digital images
can incorrectly represent the important
morphological characters of preserved
specimens.
Acknowledgements
We are grateful to Eugene Chervony and
Oleksandr Hryvul for providing their
cameras for testing, and all other people
assisting in the experiment. We are also
grateful to Patricia Mergen, Sofie De Smedt,
and Maarten Strack for consulting and
advising on the photosystems and image
optimization.
References
Baratè, A., Caccianiga, M., Caporali, E.,
Ludovico, L.A., Pinto, S., Presti, G., Sala, E., &
Testa, A. (2020). Preserving and promoting the
herbarium of the University of Milan through
digital technologies. IOP Conference Series: Materials
Science and Engineering, 949(1), Article 012066.
https://doi.org/10.1088/1757-899X/949/1/012066
Carta, A., Fernández-Pascual, E., Gioria, M.,
Müller, J.V., Rivière, S., Rosbakh, S.,
Saatkamp, A., Vandelook, F. & Mattana, E.
(2022). Climate shapes the seed germination
niche of temperate flowering plants: a meta-
analysis of European seed conservation data.
Annals of Botany, 129(7), 775–786. https://doi.
org/10.1093/aob/mcac037
https://doi.org/10.1088/1757-899X/949/1/012066
https://doi.org/10.1093/aob/mcac037
https://doi.org/10.1093/aob/mcac037
Plant Introduction • 99/100 45
Budget photosystems for the herbarium digitization purposes
Davis, C.C. (2023). The herbarium of the future.
Trends in Ecology & Evolution, 38(5), 412–423.
https://doi.org/10.1016/j.tree.2022.11.015
Dawson-Glass, E., & Hargreaves, A.L. (2022).
Does pollen limitation limit plant ranges?
Evidence and implications. Philosophical
Transactions of the Royal Society B: Biological
Sciences, 377(1846), Article 20210014. https://
doi.org/10.1098/rstb.2021.0014
Drew, J.A., Moreau, C.S., & Stiassny, M.L. (2017).
Digitization of museum collections holds the
potential to enhance researcher diversity.
Nature Ecology & Evolution, 1(12), 1789–1790.
https://doi.org/10.1038/s41559-017-0401-6
DT Heritage. (2023). Overview of FADGI &
METAMORFOZE. DT Heritage. https://heritage-
digitaltransitions.com/digitization-program-
planning/overview-of-fadgi-metamorfoze-
guidelines/
Gries, C., Gilbert, M.E.E., & Franz, N.M. (2014).
Symbiota – a virtual platform for creating
voucher-based biodiversity information
communities. Biodiversity Data Journal, 2, Article
e1114. https://doi.org/10.3897/BDJ.2.e1114
Gu, K., Liu, H., & Zhou, C. (2022). Quality
assessment of enhanced images. In K. Gu,
H. Liu, & C. Zhou, Quality Assessment of Visual
Content (pp. 127–163). Springer Singapore.
https://doi.org/10.1007/978-981-19-3347-9_5
Hammer, Ø., Harper, D.A.T., & Ryan, P.D. (2001).
PAST: Paleontological statistics software
package for education and data analysis.
Palaeontologia Electronica, 4(1), Article 4.
http://palaeo-electronica.org/2001_1/past/
issue1_01.htm
Harris, K.M., & Marsico, T.D. (2017). Digitizing
specimens in a small herbarium: a viable
workflow for collections working with limited
resources. Applications in Plant Sciences, 5(4),
Article 1600125. https://doi.org/10.3732/
apps.1600125
Hasler, D., & Suesstrunk, S. (2003).
Measuring color fulness in natural images.
Proceedings of SPIE, 5007, 87–95. https://doi.
org/10.1117/12.477378
Hedrick, B.P., Heberling, J.M., Meineke, E.K.,
Turner, K.G., Grassa, C.J., Park, D.S.,
Kennedy, J., Clarke, J.A., Cook, J.A.,
Blackburn, D.C., Edwards, S.V., & Davis, C.C.
(2020). Digitization and the future of natural
history collections. BioScience, 70(3), 243–251.
https://doi.org/10.1093/biosci/biz163
Hussein, B.R., Malik, O.A., Ong, W.-H., &
Slik, J.W.F. (2022). Applications of computer
vision and machine learning techniques for
digitized herbarium specimens: a systematic
literature review. Ecological Informatics, 69,
Article 101641. https://doi.org/10.1016/j.
ecoinf.2022.101641
iDigBio. (2014). iDigBio imaging equipment
recommendations. Ver. 2.0. iDigBio White
Papers. University of Florida. https://www.
idigbio.org/wiki/images/8/86/IDigBioImaging
GeneralEquipmentRecommendations1_0.pdf
Jackowiak, B., Lawenda, M., Nowak, M.M.,
Wolniewicz, P., Błoszyk, J., Urbaniak, M.,
Szkudlarz, P., Jędrasiak, D., Wiland-
Szymańska, J., Bajaczyk, R., & Meyer, N.
(2022). Open access to the digital biodiversity
database: a comprehensive functional model
of the natural history collections. Diversity,
14(8), Article 596. https://doi.org/10.3390/
d14080596
Khan, N., Thelwall, M., & Kousha, K. (2021).
Measuring the impact of biodiversity
datasets: data reuse, citations and altmetrics.
Scientometrics, 126, 3621–3639. https://doi.
org/10.1007/s11192-021-03890-6
Le Page, M. (2022, May 19). Priceless samples
from Ukraine’s seed bank destroyed in bomb
attack. New Scientist. https://www.newscientist.
com/art ic le/2321008-priceless-samples-
from-ukraines-seed-bank-destroyed- in-
bomb-attack/
Mosyakin, S.L., & Shiyan, N.M. (2022). The
M.G. Kholodny Institute of Botany and the
National Herbarium of Ukraine (KW), Kyiv:
Damage due to the missile strikes on 10
October 2022. Ukrainian Botanical Journal,
79(5), 339–342. https://ukrbotj.co.ua/
pdf/79/5/ukrbotj-2022-79-5-339.pdf
Nelson, G. & Ellis, S. (2019). The history and
impact of digitization and digital data
mobilization on biodiversity research.
Philosophical Transactions of the Royal Society B:
Biological Sciences, 374(1763), Article 20170391.
https://doi.org/10.1098/rstb.2017.0391
Nieva de la Hidalga, A., Rosin, P.L., Sun, X.,
Bogaerts, A., De Meeter, N., De Smedt, S.,
Strack van Schijndel, M., Van Wambeke, P.,
& Groom, Q. (2020). Designing an herbarium
digitisation workflow with built-in image
quality management. Biodiversity Data Journal,
8, Article e47051. https://doi.org/10.3897/
BDJ.8.e47051
Novikov, A. (2019, November 27–29). Digitization
of natural collections – the way to immortality.
In Abstracts of the 14th International Young
Scientists’ Conference “Biology: From a Molecule
Up to the Biosphere” (pp. 12–14). V.N. Karazin
Kharkiv National University. https://doi.
org/10.5281/zenodo.3594474
https://doi.org/10.1016/j.tree.2022.11.015
https://doi.org/10.1098/rstb.2021.0014
https://doi.org/10.1098/rstb.2021.0014
https://doi.org/10.1038/s41559-017-0401-6
https://heritage-digitaltransitions.com/digitization-program-planning/overview-of-fadgi-metamorfoze-
https://heritage-digitaltransitions.com/digitization-program-planning/overview-of-fadgi-metamorfoze-
https://heritage-digitaltransitions.com/digitization-program-planning/overview-of-fadgi-metamorfoze-
https://heritage-digitaltransitions.com/digitization-program-planning/overview-of-fadgi-metamorfoze-
https://doi.org/10.3897/BDJ.2.e1114
https://doi.org/10.1007/978-981-19-3347-9_5
http://palaeo-electronica.org/2001_1/past/issue1_01.htm
http://palaeo-electronica.org/2001_1/past/issue1_01.htm
https://doi.org/10.3732/apps.1600125
https://doi.org/10.3732/apps.1600125
https://doi.org/10.1117/12.477378
https://doi.org/10.1117/12.477378
https://doi.org/10.1093/biosci/biz163
https://doi.org/10.1016/j.ecoinf.2022.101641
https://doi.org/10.1016/j.ecoinf.2022.101641
https://www.idigbio.org/wiki/images/8/86/IDigBioImagingGeneralEquipmentRecommendations1_0.pdf
https://www.idigbio.org/wiki/images/8/86/IDigBioImagingGeneralEquipmentRecommendations1_0.pdf
https://www.idigbio.org/wiki/images/8/86/IDigBioImagingGeneralEquipmentRecommendations1_0.pdf
https://doi.org/10.3390/d14080596
https://doi.org/10.3390/d14080596
https://doi.org/10.1007/s11192-021-03890-6
https://doi.org/10.1007/s11192-021-03890-6
https://www.newscientist.com/article/2321008-priceless-samples-from-ukraines-seed-bank-destroyed-in-
https://www.newscientist.com/article/2321008-priceless-samples-from-ukraines-seed-bank-destroyed-in-
https://www.newscientist.com/article/2321008-priceless-samples-from-ukraines-seed-bank-destroyed-in-
https://www.newscientist.com/article/2321008-priceless-samples-from-ukraines-seed-bank-destroyed-in-
https://ukrbotj.co.ua/pdf/79/5/ukrbotj-2022-79-5-339.pdf
https://ukrbotj.co.ua/pdf/79/5/ukrbotj-2022-79-5-339.pdf
https://doi.org/10.1098/rstb.2017.0391
https://doi.org/10.3897/BDJ.8.e47051
https://doi.org/10.3897/BDJ.8.e47051
https://doi.org/10.5281/zenodo.3594474
https://doi.org/10.5281/zenodo.3594474
46 Plant Introduction • 99/100
Novikov et al.
Novikov, A., & Sup-Novikova, M. (2021). Simple
and cheap photosystem for herbarium
digitization. Plant Introduction, 91/92, 50–53.
https://doi.org/10.46341/PI2021015
Novikov, A.V., Hushtan, H.H.,Hushtan, K.V.,
Kuzyarin, O.T., Leleka, D.Y., Nachychko, V.O.,
Prots, B.H., Rizun, V.B., Savytska, A.G.,
Susulovska, S.A., & Susulovsky, A.S. (2023).
Outlining the aims and format of the project
“Digitisation of natural history collections
damaged as a result of hostilities and related
factors: development of protocols and
implementation based on the State Museum
of Natural History of the National Academy
of Sciences of Ukraine”. Scientific Proceedings
of the State Museum of Natural History of the
National Academy of Sciences of Ukraine, 39, 19–
30. (In Ukrainian). https://doi.org/10.36885/
nzdpm.2023.39.19-30
Palus, H. (2006). Colorfulness of the image:
definition, computation, and properties.
Proceedings of SPIE, 6158, Article 615805.
https://doi.org/10.1117/12.675760
Pavlyshyn, O. (2022, June 01). What happened
to the only Ukrainian genetic bank of plants.
Kunsht. (In Ukrainian). https://kunsht.com.ua/
articles/shho-naspravdi-stalosya-z-yedinim-
v-ukraini-genetichnim-bankom-roslin
Picturae. (2023). Delt.ae. Picturae. https://deltae.
picturae.com/
Rieger, T. (Ed.). (2016). Technical guidelines for
digitizing cultural heritage materials: creation of
raster image files. Federal Agencies Digitization
Guidelines Initiative, USA. https://www.
d i g i t i z a t i o n g u i d e l i n e s . g o v / g u i d e l i n e s /
FADGI%20Federa l%20%20Agenc ies %20
Digital%20Guidelines%20Initiative-2016%20
Final_rev1.pdf
Rieger, T., Phelps, K.A., Beckerle, H.,
Brown, T., Frederick, R., Mitrani, S.,
Breen, P., Breitbart, M., Williams, D.,
Triplett, R., & Horsley, M. (2023). Technical
guidelines for digitizing cultural heritage
materials (3rd ed.). Federal Agencies
Digitization Guidelines Initiative, USA.
https:/ /www.dig i t izat ionguidel ines.gov/
g u i d e l i n e s / F A D G I % 2 0 T e c h n i c a l % 2 0
G u i d e l i n e s % 2 0 f o r % 2 0 D i g i t i z i n g % 2 0
Cultural%20Heritage%20Materials_3rd%20
Edition_05092023.pdf
Samain, M.-S., Guzmán Díaz, S., Machuca
Machuca, K., Dolores Fuentes, A.C., Zacarías
Correa, A.G., Valentín Martínez, D., Aldaba
Núñez, F.A., Redonda-Martínez, R., Oldfield, S.F.,
& Martínez Salas, E.M. (2023). Meta-analysis of Red
List conservation assessments of Mexican endemic
and near endemic tree species shows nearly two
thirds of these are threatened. Plants, People, Planet,
5(4), 581–599. https://doi.org/10.1002/ppp3.10308
Schindel, D.E., & Cook, J.A. (2018). The next
generation of natural history collections.
PLOS Biology, 16(7), Article e2006125. https://
doi.org/10.1371/journal.pbio.2006125
Takano, A., Horiuchi, Y., Fujimoto, Y., Aoki, K.,
Mitsuhashi, H., & Takahashi, A. (2019).
Simple but long-lasting: a specimen imaging
method applicable for small- and medium-
sized herbaria. PhytoKeys, 118, 1–14. https://
doi.org/10.3897/phytokeys.118.29434
Thiers, B.M. (2023). The world’s herbaria 2022:
A summary report based on data from Index
Herbariorum, Issue 6.0. New York Botanical
Garden. https://sweetgum.nybg.org/science/
wp-content/uploads/2023/10/The_Worlds_
Herbaria_2022_Report.docx
Vajda, V., McLoughlin, S., & Shevchuk, O. (2022,
June 19–20). The war in Ukraine – Its impact
on palaeobotany, palynology, herbaria and
museums. In S. McLoughlin (Ed.), 11th European
Palaeobotany and Palynology Conference
Abstracts, Program and Proceedings (pp. 75–76).
Swedish Museum of Natural History.
van Dormolen, H. (2012). Metamorfoze preservation
imaging guidelines: Image quality, version 1.0,
January 2012. Koninklijke Bibliotheek. https://
www.metamorfoze.nl/sites/default/f i les/
documents/Metamorfoze_Preservat ion_
Imaging_Guidelines_1.0.pdf
van Dormolen, H. (2019, May 14–17).
Metamorfoze preservation imaging
guidelines, version 2.0. In Proceedings of the IS&T
Conference “Archiving 2019” (pp. 9–11). Society
for Imaging Science and Technology. https://
doi.org/10.2352/issn.2168-3204.2019.1.0.3
Zerman, E., Rana, A., & Smolic, A. (2019, September
22–25). Colornet – estimating colorfulness in natural
images. In Proceedings of the 2019 IEEE International
Conference on Image Processing (ICIP) (pp. 3791–3795).
Institute of Electrical and Electronics Engineers
(IEEE). https://doi.org/10.1109/ICIP.2019.8803407
https://doi.org/10.46341/PI2021015
https://doi.org/10.36885/nzdpm.2023.39.19-30
https://doi.org/10.36885/nzdpm.2023.39.19-30
https://doi.org/10.1117/12.675760
https://kunsht.com.ua/articles/shho-naspravdi-stalosya-z-yedinim-v-ukraini-genetichnim-bankom-roslin
https://kunsht.com.ua/articles/shho-naspravdi-stalosya-z-yedinim-v-ukraini-genetichnim-bankom-roslin
https://kunsht.com.ua/articles/shho-naspravdi-stalosya-z-yedinim-v-ukraini-genetichnim-bankom-roslin
https://deltae.picturae.com/
https://deltae.picturae.com/
https://www.digitizationguidelines.gov/guidelines/FADGI%20Federal%20%20Agencies%20Digital%20Guidelin
https://www.digitizationguidelines.gov/guidelines/FADGI%20Federal%20%20Agencies%20Digital%20Guidelin
https://www.digitizationguidelines.gov/guidelines/FADGI%20Federal%20%20Agencies%20Digital%20Guidelin
https://www.digitizationguidelines.gov/guidelines/FADGI%20Federal%20%20Agencies%20Digital%20Guidelin
https://www.digitizationguidelines.gov/guidelines/FADGI%20Federal%20%20Agencies%20Digital%20Guidelin
https://www.digitizationguidelines.gov/guidelines/FADGI%20Technical%20Guidelines%20for%20Digitizing%
https://www.digitizationguidelines.gov/guidelines/FADGI%20Technical%20Guidelines%20for%20Digitizing%
https://www.digitizationguidelines.gov/guidelines/FADGI%20Technical%20Guidelines%20for%20Digitizing%
https://www.digitizationguidelines.gov/guidelines/FADGI%20Technical%20Guidelines%20for%20Digitizing%
https://www.digitizationguidelines.gov/guidelines/FADGI%20Technical%20Guidelines%20for%20Digitizing%
https://doi.org/10.1002/ppp3.10308
https://doi.org/10.1371/journal.pbio.2006125
https://doi.org/10.1371/journal.pbio.2006125
https://doi.org/10.3897/phytokeys.118.29434
https://doi.org/10.3897/phytokeys.118.29434
https://sweetgum.nybg.org/science/wp-content/uploads/2023/10/The_Worlds_Herbaria_2022_Report.docx
https://sweetgum.nybg.org/science/wp-content/uploads/2023/10/The_Worlds_Herbaria_2022_Report.docx
https://sweetgum.nybg.org/science/wp-content/uploads/2023/10/The_Worlds_Herbaria_2022_Report.docx
https://www.metamorfoze.nl/sites/default/files/documents/Metamorfoze_Preservation_Imaging_Guidelines
https://www.metamorfoze.nl/sites/default/files/documents/Metamorfoze_Preservation_Imaging_Guidelines
https://www.metamorfoze.nl/sites/default/files/documents/Metamorfoze_Preservation_Imaging_Guidelines
https://www.metamorfoze.nl/sites/default/files/documents/Metamorfoze_Preservation_Imaging_Guidelines
https://doi.org/10.2352/issn.2168-3204.2019.1.0.3
https://doi.org/10.2352/issn.2168-3204.2019.1.0.3
https://doi.org/10.1109/ICIP.2019.8803407
Plant Introduction • 99/100 47
Budget photosystems for the herbarium digitization purposes
Appendix A. Tested ∆E2000 values for images provided by different virtual herbaria.
Herbarium
acronym
LWS BR L P
Data source http://dc.smnh.org/ https://www.
botanicalcollections.be
https://bioportal.
naturalis.nl/
https://science.mnhn.
fr/institution/mnhn/
collection/p/list
Image
format
jpeg jpeg jpeg jpeg
Specimen’s
image
∆E2000 Specimen’s
image
∆E2000 Specimen’s
image
∆E2000 Specimen’s
image
∆E2000
Tested
images
LWS114876 7.27 BR0000005610143 4.44 L2706121 1.77 P05606059 7.91
LWS115814 7.13 BR0000005251551 8.07 L4227700 2.58 P05604378 7.65
LWS110057 8.98 BR0000005085590 6.56 L1824123 1.65 P05556372 11.94
LWS108434 7.28 BR0000005268429 9.71 L1824121 1.61 P06616638 10.98
LWS112356 8.58 BR0000005269457 11.55 L1824100 1.77 P06614510 4.67
LWS114666 6.83 BR0000005268962 10.05 L1664736 1.5 P06665729 4.76
LWS064350 7.09 BR0000005599288 6.45 L1664752 1.53 P06664725 3.98
LWS012269 7.07 BR0000005574506 8.96 L2048207 1.74 P06643009 5.78
LWS074085 8.79 BR0000005610044 4.61 L1664728 1.68 P06872198 9.49
LWS097168 7.69 BR0000005610129 4.4 L1664518 1.63 P06898681 10.03
LWS093189 9.98 BR0000005612376 3.99 L4201532 2.75 P05572674 4.16
LWS104671 7.54 BR0000005610310 4.75 L4206155 2.52 P06169913 10.41
LWS082301 6.46 BR0000005647286 5.38 L3912403 2.06 P05348062 4.85
LWS090820 7.34 BR0000005692408 2.93 L3880536 2.11 P05349490 9.55
LWS090821 8.92 BR0000005693276 3.84 L3798683 2.48 P05349809 8.89
LWS092529 8.41 BR0000005756421 8.18 L3421811 2.43 P05352452 11.96
LWS095436 6.52 BR0000005749218 8.94 L3422024 2.57 P05413635 9.93
LWS117591 6.69 BR0000005751273 8.42 L3798348 2.35 P05401397 10.61
LWS041914 8.46 BR0000006010706 7.54 L1826712 1.66 P00114597 5.73
LWS041798 6.62 BR0000006014834 7.56 L1826706 1.61 P05325382 10.48
LWS104998 8.71 BR0000006397081 7.75 L1729042 1.55 P05330196 10.47
LWS045196 7.62 BR0000006386917 6.7 L1726296 1.59 P05377102 8.15
LWS081841 7.53 BR0000006445683 7.64 L1712761 1.67 P05345786 10.06
LWS111288 9.42 BR0000006468316 6.42 L1714522 1.87 P05356212 10.97
LWS117303 6.92 BR0000006651022 7.7 L1229311 1.45 P05338196 6.28
LWS007193 6.78 BR0000009737181 7.16 L3017598 1.86 P05338210 7.54
LWS007104 7.4 BR0000011661016 5.15 L3012534 2.18 P00102771 8.4
LWS116675 6.44 BR0000008547156 10.62 L4172147 2.1 P00102794 4.87
LWS059278 7.45 BR0000019150062 2.79 L2048207 1.74 P00095479 6.28
LWS059277 7.48 BR0000028193302 7.49 L1723201 2.05 P00455987 4.97
Excluded
outliers
LWS118511 4.71 L4307171 10.22 P05094218 14.98
LWS117185 4.9 L2070825 8.57
L1664519 3.9
L3299057 3.55
http://dc.smnh.org/
https://www.botanicalcollections.be
https://www.botanicalcollections.be
https://bioportal.naturalis.nl/
https://bioportal.naturalis.nl/
https://science.mnhn.fr/institution/mnhn/collection/p/list
https://science.mnhn.fr/institution/mnhn/collection/p/list
https://science.mnhn.fr/institution/mnhn/collection/p/list
48 Plant Introduction • 99/100
Novikov et al.
Appendix A. Continued.
Herbarium
acronym
B US NY W
Data source https://herbarium.gbif.
de/
https://collections.nmnh.
si.edu/search/botany/
https://sweetgum.nybg.
org/science/vh/
https://www.jacq.org/
Image
format
jpeg jpeg jpeg tiff
Specimen’s
image
∆E2000 Specimen’s
image
∆E2000 Specimen’s
image
∆E2000 Specimen’s
image
∆E2000
Tested
images
B10 0612505 11.64 US02657192 5.71 NY01154240 5.13 W0022412 9.08
B10 0762104 13.29 US02036841 5.7 NY00659014 4.06 W20180011081 9.5
B10 0463047 10.95 US01638693 6.72 NY00278673 4.47 W19260002399 6.95
B10 0720818 12.7 US02295206 5.72 NY01111411 3.62 W0198224 5.87
B10 0295345 8.83 US02295544 5.81 NY01207928 3.56 W20150015230 9.53
B10 0062990 13.67 US02295466 5.55 NY00380615 4.16 W20040015205 7.93
B10 0498955 10.66 US02299047 6.32 NY01041711 4.4 W19940007184 10.12
B10 0194679 11.76 US01236911 3.23 NY01240070 4.91 W19940005381 9.43
B10 0501215 11.45 US02910399 6.48 NY01240069 4.6 W19750014148 8.2
B10 0401925 12.48 US03315596 6.41 NY00429915 5.71 W19710021586 8.59
BW15774010 6.59 US03558373 5.22 NY00579492 4.2 W19730010518 10.22
B10 0580201 14.33 US03648212 3.91 NY00429301 4.17 W18890358074 7.93
BW09847010 10.93 US03655213 4.27 NY00415855 4.41 W19160040971 7.06
BW09840030 9.75 US03708428 4.3 NY00429300 4.14 W19110006601 16.39
BW09846020 10.39 US03708743 3.84 NY00415847 4.42 W18890344142 6.36
B100356299 5.73 US03965306 3.66 NY00415854 4.23 W18890343339 6.07
B100628095 11.66 US00498975 5.55 NY00381397 4.96 W18890176850 7.19
BW00879010 8.09 US01341530 4.56 NY00353423 5.55 W18890169647 15.49
BW00882010 8.3 US01351765 5.63 NY00353390 5.64 W18890116066 11.12
B100626453 15.62 US01418103 7.03 NY00353343 5.26 W18890167221 10.33
B100463256 14.59 US01418118 6.84 NY00353304 5.59 W18890103497 16.35
BW17516010 8.13 US01638679 7.04 NY00353320 5.72 W18890105101 14.94
BW18366010 7.3 US03034171 4.66 NY00335712 4.25 W18890078203 14.94
BW18366040 7.95 US01562628 6.26 NY00250998 4.54 W0074239 12.74
B100279026 9.41 US00371149 5.34 NY00232257 4.42 W0209299 6.14
B100277369 10.6 US00315479 3.81 NY00180101 4.25 W0205862 5.35
B100576703 13.12 US03910541 4.58 NY00074332 4.49 W0031990 6.09
B100629675 13.3 US03910533 4.3 NY00162932 4.08 W0064125 8.96
B101041345 13.85 US03825241 4.6 NY00025829 4.28 W0017426 15.84
BW06535010 10.23 US03825258 4.56 NY00004050 5.29 W0022052 5.64
Excluded
outliers
US03018521 18.7 NY00709539 7.39
NY01081819 9.96
NY01082967 8.71
https://herbarium.gbif.de/
https://herbarium.gbif.de/
https://collections.nmnh.si.edu/search/botany/
https://collections.nmnh.si.edu/search/botany/
https://sweetgum.nybg.org/science/vh/
https://sweetgum.nybg.org/science/vh/
https://www.jacq.org/
Plant Introduction • 99/100 49
Budget photosystems for the herbarium digitization purposes
Herbarium
acronym
BM BM GZU PI
Data source https://data.nhm.ac.uk/ https://data.nhm.ac.uk/ https://www.jacq.org/ https://www.jacq.org/
Image
format
tiff jpeg tiff tiff
Specimen’s
image
∆E2000 Specimen’s
image
∆E2000 Specimen’s
image
∆E2000 Specimen’s
image
∆E2000
Tested
images
BM001046253 5.87 BM000521903 6.55 GZU000312195 5.46 PI018084 6.46
BM001134949 6.95 BM000051636 6.41 GZU000330246 5.72 PI023399 10.08
BM001125317 5.91 BM000051651 7.04 GZU000302173 6.03 PI016605 6.69
BM001043579 4.58 BM000522332 6.5 GZU000302171 5.64 PI010679 5.97
BM001042230 6.19 BM001024837 6.14 GZU000302165 5.59 PI015062 6.09
BM001009514 5.79 BM001024838 6.14 GZU000302164 6.05 PI015053 5.97
BM000646123 5.89 BM000946954 9.6 GZU000302124 6.03 PI010503 6.37
BM001008620 6.38 BM000946956 9.08 GZU000302114 4.35 PI060322 7.82
BM000885992 5.61 BM000946961 9.78 GZU000302105 6.53 PI062216 8.02
BM000646237 6.03 BM000936510 8.3 GZU000302045 4.71 PI058931 10.02
BM000040272 4.48 BM000936516 8.44 GZU000299033 5.15 PI058227 8.32
BM000557808 5.11 BM000829164 8.83 GZU000302036 4.8 PI054634 9
BM000646120 5.57 BM000939392 9.36 GZU000299772 5.78 PI055889 9.47
BM000557807 5.18 BM000936518 8 GZU000294923 6.24 PI054590 8.23
BM000557806 5.47 BM001042975 8.89 GZU000277836 4.37 PI051599 8.57
BM000557803 5.03 BM001072227 9.42 GZU000294921 6.26 PI052647 6.94
BM000557664 5.2 BM001072361 9.31 GZU000273692 5.19 PI051525 8.27
BM000051695 6.38 BM001024842 6.1 GZU000277834 4.4 PI051558 8.53
BM000051628 5.29 BM000062783 9.37 GZU000273469 5.17 PI051418 7.93
BM000042611 6.25 BM000939063 8.95 GZU000260591 5.08 PI047655 9.52
BM000051154 5.67 BM000953182 8.75 GZU000259529 5.51 PI047640 10.17
BM000051570 6.7 BM000997729 9.33 GZU000259696 4.27 PI043447 11.3
BM001057204 7.02 BM000997734 9.26 GZU000250661 5.05 PI043446 10.78
BM001067297 6.39 BM000041699 9.38 GZU000279097 5.17 PI043154 9.45
BM000954795 4.72 BM000042712 9.4 GZU000277963 5.16 PI035023 9.57
BM001042138 4.41 BM000939451 8.81 GZU000250527 5.47 PI033791 8.51
BM000938607 5.28 BM000939449 9.71 GZU000120870 4.56 PI040569 9.26
BM000646067 6.2 BM000946986 9.17 GZU000249376 4.75 PI021697 10.98
BM000042227 4.36 BM001024149 9.05 GZU000093404 5.23 PI030571 9.54
BM000609397 6.24 BM001024147 9.36 GZU000249715 5.5 PI030557 8.56
Excluded
outliers
BM001134966 10.83 BM000895849 12.85 GZU000251150 10.03
BM000589029 11.89 BM000895855 13.38 GZU000251754 10.74
BM000621852 9.42 BM000895858 11.62 GZU000250659 9.35
BM001008628 3.76 BM000895860 12.66 GZU000249458 8.46
BM000042612 11.64
BM000946985 10.41
Appendix A. Continued.
https://data.nhm.ac.uk/
https://data.nhm.ac.uk/
https://www.jacq.org/
https://www.jacq.org/
50 Plant Introduction • 99/100
Novikov et al.
Тестування бюджетних фотосистем з метою досягнення оптимального
результату для цілей оцифрування гербарію
Андрій Новіков 1, *, Марія Суп-Новікова 2, Віктор Начичко 3, Олександр Кузярін 1
1 Державний природознавчий музей НАН України, вул. Театральна, 18, Львів, 79008, Україна;
* novikoffav@gmail.com
2 Український католицький університет, вул. Стрийська, 29A, Львів, 79026, Україна
3 Львівський національний університет імені Івана Франка, вул. Грушевського, 4, Львів, 79005,
Україна
Ця робота стосується питання застосування різних бюджетних фотосистем для оцифрування
гербарних зразків. Було випробувано дванадцять фотосистем щодо точності кольоропередачі
зображень, які вони продукують. Встановлено, що фотосистема, побудована на базі фотоапарата
Canon EOS 800D та фіксованого об’єктива Tokina AT-X M35 PRO DX AF 35 mm f/2.8 Macro, яка зараз
використовується для оцифрування гербарної колекції LWS, показала найкращі результати серед
інших протестованих фотосистеми. Ця фотосистема також продукувала зображення з такою ж або
навіть кращою точністю кольоропередачі, як і зображення, завантажені з віртуальних гербаріїв P,
PI, B і W. Тим не менш, точність кольоропередачі, виміряна як ΔE2000, загалом не відповідає останнім
критеріям, встановленим для цілей оцифрування гербарію та новим вимогам FADGI. Було виявлено,
що ця фотосистема має потенціал для вдосконалення, а отже, її слід оптимізувати. З іншого боку,
також було показано, що смартфон Samsung Galaxy S10 може створювати зображення з такою ж
або навіть кращою точністю кольоропередачі порівняно з деякими зображеннями, збереженими
у віртуальних гербаріях P, PI, W і B. Тому, в умовах екстремальних ситуацій та бойових дій такі
смартфони з додатковим зовнішнім освітленням можуть слугувати для екстренного оцифрування
природничих колекцій. Поза тим, ми сумніваємося у необхідності застосування загальновживаних
кольорових мішеней для оцифрування гербарних зразків, оскільки оригінальний колір таких зразків
зазвичай втрачається під час гербаризації та збереження. Замість цього було б більш доцільно
застосовувати розширені мішені для оцінки просторової точності зображень, оскільки зображення
можуть спотворено відображати важливі морфологічні характеристики таких зразків.
Ключові слова: оцифрування гербарію, оцінка якості зображень, фотосистеми, дельта E, точність кольоропередачі
|
| id | oai:ojs2.plantintroduction.org:article-1631 |
| institution | Plant Introduction |
| keywords_txt_mv | keywords |
| language | English |
| last_indexed | 2025-07-17T12:54:18Z |
| publishDate | 2023 |
| publisher | M.M. Gryshko National Botanical Garden of the NAS of Ukraine |
| record_format | ojs |
| resource_txt_mv | wwwplantintroductionorg/f9/90661450413f6bb2e6cdea1a42db35f9.pdf |
| spelling | oai:ojs2.plantintroduction.org:article-16312024-04-07T19:57:15Z Testing budget photosystems to reach an optimal solution for the herbarium digitization purposes Тестування бюджетних фотосистем з метою досягнення оптимального результату для цілей оцифрування гербарію Novikov, Andriy Sup-Novikova, Mariia Nachychko, Viktor Kuzyarin, Oleksandr The current paper stresses the application of different budget photosystems for digitization of herbarium specimens. Twelve photosystems were compared by color accuracy reproduction of the images. It was found that the photosystem built on the basis of photocamera Canon EOS 800D and fixed lens Tokina AT-X M35 PRO DX AF 35 mm f/2.8 Macro and currently used for the digitization of the LWS herbarium collection demonstrated the best results among other tested photosystems. It also produced the images with the same or even better color accuracy as in images downloaded from P, PI, B, and W virtual herbaria. Nevertheless, its color accuracy measured as ΔE2000, in general, does not meet recent criteria ascertained for the herbarium digitization purposes and new FADGI’s requirements. This photosystem has been found to have improving potential and, hence, should be optimized. On the other hand, it was also shown that smartphone Samsung Galaxy S10 could produce images with the same or even better color accuracy compared to some images deposited at P, PI, W, and B virtual herbaria. Therefore, in conditions of extreme situations and hostilities, such smartphones with additional external illumination can serve for urgent digitization of natural history collections. Finally, we doubt the application of commonly used color targets for the digitization of herbariaum preserved specimens since the original living color of such specimens is usually lost during conservation and preservation. Instead, it would be more beneficial to apply advanced targets to evaluate the spatial accuracy of images since they can incorrectly represent the important morphological characters of preserved specimens. Ця робота стосується питання застосування різних бюджетних фотосистем для оцифрування гербарних зразків. Було випробувано дванадцять фотосистем щодо точності кольоропередачі зображень, які вони продукують. Встановлено, що фотосистема, побудована на базі фотоапарата Canon EOS 800D та фіксованого об’єктива Tokina AT-X M35 PRO DX AF 35 mm f/2.8 Macro, яка зараз використовується для оцифрування гербарної колекції LWS, показала найкращі результати серед інших протестованих фотосистеми. Ця фотосистема також продукувала зображення з такою ж або навіть кращою точністю кольоропередачі, як і зображення, завантажені з віртуальних гербаріїв P, PI, B і W. Тим не менш, точність кольоропередачі, виміряна як ΔE2000, загалом не відповідає останнім критеріям, встановленим для цілей оцифрування гербарію та новим вимогам FADGI. Було виявлено, що ця фотосистема має потенціал для вдосконалення, а отже, її слід оптимізувати. З іншого боку, також було показано, що смартфон Samsung Galaxy S10 може створювати зображення з такою ж або навіть кращою точністю кольоропередачі порівняно з деякими зображеннями, збереженими у віртуальних гербаріях P, PI, W і B. Тому, в умовах екстремальних ситуацій та бойових дій такі смартфони з додатковим зовнішнім освітленням можуть слугувати для екстренного оцифрування природничих колекцій. Поза тим, ми сумніваємося у необхідності застосування загальновживаних кольорових мішеней для оцифрування гербарних зразків, оскільки оригінальний колір таких зразків зазвичай втрачається під час гербаризації та збереження. Замість цього було б більш доцільно застосовувати розширені мішені для оцінки просторової точності зображень, оскільки зображення можуть спотворено відображати важливі морфологічні характеристики таких зразків. M.M. Gryshko National Botanical Garden of the NAS of Ukraine 2023-11-24 Article Article application/pdf https://www.plantintroduction.org/index.php/pi/article/view/1631 10.46341/PI2023010 Plant Introduction; No 99/100 (2023); 36-50 Інтродукція Рослин; № 99/100 (2023); 36-50 2663-290X 1605-6574 10.46341/PI99-100 en https://www.plantintroduction.org/index.php/pi/article/view/1631/1547 Copyright (c) 2023 Andriy Novikov, Mariia Sup-Novikova, Viktor Nachychko, Oleksandr Kuzyarin http://creativecommons.org/licenses/by/4.0 |
| spellingShingle | Novikov, Andriy Sup-Novikova, Mariia Nachychko, Viktor Kuzyarin, Oleksandr Тестування бюджетних фотосистем з метою досягнення оптимального результату для цілей оцифрування гербарію |
| title | Тестування бюджетних фотосистем з метою досягнення оптимального результату для цілей оцифрування гербарію |
| title_alt | Testing budget photosystems to reach an optimal solution for the herbarium digitization purposes |
| title_full | Тестування бюджетних фотосистем з метою досягнення оптимального результату для цілей оцифрування гербарію |
| title_fullStr | Тестування бюджетних фотосистем з метою досягнення оптимального результату для цілей оцифрування гербарію |
| title_full_unstemmed | Тестування бюджетних фотосистем з метою досягнення оптимального результату для цілей оцифрування гербарію |
| title_short | Тестування бюджетних фотосистем з метою досягнення оптимального результату для цілей оцифрування гербарію |
| title_sort | тестування бюджетних фотосистем з метою досягнення оптимального результату для цілей оцифрування гербарію |
| url | https://www.plantintroduction.org/index.php/pi/article/view/1631 |
| work_keys_str_mv | AT novikovandriy testingbudgetphotosystemstoreachanoptimalsolutionfortheherbariumdigitizationpurposes AT supnovikovamariia testingbudgetphotosystemstoreachanoptimalsolutionfortheherbariumdigitizationpurposes AT nachychkoviktor testingbudgetphotosystemstoreachanoptimalsolutionfortheherbariumdigitizationpurposes AT kuzyarinoleksandr testingbudgetphotosystemstoreachanoptimalsolutionfortheherbariumdigitizationpurposes AT novikovandriy testuvannâbûdžetnihfotosistemzmetoûdosâgnennâoptimalʹnogorezulʹtatudlâcílejocifruvannâgerbaríû AT supnovikovamariia testuvannâbûdžetnihfotosistemzmetoûdosâgnennâoptimalʹnogorezulʹtatudlâcílejocifruvannâgerbaríû AT nachychkoviktor testuvannâbûdžetnihfotosistemzmetoûdosâgnennâoptimalʹnogorezulʹtatudlâcílejocifruvannâgerbaríû AT kuzyarinoleksandr testuvannâbûdžetnihfotosistemzmetoûdosâgnennâoptimalʹnogorezulʹtatudlâcílejocifruvannâgerbaríû |