9762 Vol. 56, No.

35 / December 10 2017 / Applied Optics Research Article

Detecting crop population growth using

chlorophyll fluorescence imaging
HENG WANG,1 XIANGJIE QIAN,2 LAN ZHANG,2 SAILONG XU,3 HAIFENG LI,1,* XIAOJIAN XIA,2
LIANKUI DAI,3 LIANG XU,1 JINGQUAN YU,2 AND XU LIU1
1
State Key Lab of Modern Optical Instrumentation, Zhejiang University, Hangzhou 310027, China
2
Department of Horticulture, Zhejiang University, Hangzhou 310058, China
3
College of Control Science and Engineering, Zhejiang University, Hangzhou 310027, China
*Corresponding author: [email protected]

Received 29 September 2017; revised 8 November 2017; accepted 8 November 2017; posted 13 November 2017 (Doc. ID 308279);
published 7 December 2017

For both field and greenhouse crops, it is challenging to evaluate their growth information on a large area over a
long time. In this work, we developed a chlorophyll fluorescence imaging-based system for crop population
growth information detection. Modular design was used to make the system provide high-intensity uniform
illumination. This system can perform modulated chlorophyll fluorescence induction kinetics measurement
and chlorophyll fluorescence parameter imaging over a large area of up to 45 cm × 34 cm. The system can
provide different lighting intensity by modulating the duty cycle of its control signal. Results of continuous mon-
itoring of cucumbers in nitrogen deficiency show the system can reduce the judge error of crop physiological
status and improve monitoring efficiency. Meanwhile, the system is promising in high throughput application
scenarios. © 2017 Optical Society of America
OCIS codes: (120.1880) Detection; (120.4640) Optical instruments; (170.2945) Illumination design; (170.6280) Spectroscopy,
fluorescence and luminescence.

https://doi.org/10.1364/AO.56.009762

1. INTRODUCTION Variation of fluorescence intensity can indicate photosynthesis

One critical aspect of agricultural production management and the protection mechanism of the plants. Chlorophyll fluo-
is continuous and nondestructive monitoring of the plant rescence detection is thus a very rich observable of photosyn-
population’s physiological status [1,2]. Despite numerous thetic activity. A few parameters can be defined and calculated
efforts [3,4], it is still challenging to do it over a long time over from chlorophyll fluorescence detection [9]: minimal fluores-
a large area. cence from dark leaf (F o ), maximal fluorescence from dark leaf
Among the currently available techniques for detecting (F m ), variable fluorescence level in the dark-adapted state (F v ),
plants’ physiological status, e.g., photosynthesis performance, maximum quantum efficiency of PSII photochemistry (F v ∕F m ),
chlorophyll fluorescence imaging has been widely used [5,6]. PSII operating efficiency (ΦPSII ), and so on.
Thanks to Kautsy and Kirsch’s work [7], the relationship Chlorophyll fluorescence can be measured with either non-
between fluorescence parameters and photosynthetic electron modulated methods [7,10] or modulated methods. Modulated
transport in vivo has been well understood: the highly complex chlorophyll fluorescence measurement has developed from
photosynthesis can be described by a simple photosystem II point detection [11] to imaging detection [12,13]. Imaging
(PSII) photochemistry model [8]. The light energy absorbed of chlorophyll fluorescence is becoming increasingly popular
by PSII can be divided in three channels. The first channel as a screening and diagnostic tool due mostly to the develop-
is photosynthesis, which accounts for 80% of the total energy ment of instrumentation [14–17], with many commercial in-
for healthy leaves of plants. Change of this channel causes struments such as Photon Systems Instruments (http://www.psi
photochemical quenching of fluorescence. The second channel .cz/), Walz (http://www.walz.com/), and Hanstech Instruments
is heat dissipation (10%), which causes nonphotochemical (http://www.hansatech-instruments.com/). In assessment of
quenching of fluorescence. Strong light is harmful to the plant sample spatial and temporal heterogeneity, the imaging tech-
and therefore heat dissipation can be a protection mechanism. nique has an inherent advantage [18]. Meanwhile, chlorophyll
The last channel is chlorophyll fluorescence, the re-emitting fluorescence imaging provides the possibility of high throughput
light is passive process affected by the former two channels. analysis.

1559-128X/17/359762-08 Journal © 2017 Optical Society of America

 Research Article Vol. 56, No. 35 / December 10 2017 / Applied Optics 9763

However, the imaging detection area is still limited, such as

10 cm × 10 cm of vivo fluorescence imaging setup designed
by Johnson et al. [19], 10 cm × 13 cm of IMAGING-PAM
of Walz (http://www.walz.com/pro-ducts/chl_p700/overview
.html), and 20 cm × 20 cm of FluorCam of Photon Systems
Instruments (http://www.psi.cz/products/fluor-cams). The de-
tection area is directly related to the amount of data obtained.
Considering the size of the crops and seedlings, 10 cm × 10 cm
detection area can only image several leaves of the plants, and
20 cm × 20 cm detection area can at most image one or two Fig. 1. Block diagram of the modulated chlorophyll fluorescence
seedlings. In this case, the physiological status of individual imaging system.
leaves or plants will interfere with the judgement of the overall
crop growing condition. There, to sample the entire crop pop-
ulation, multiple measurements are necessary, restricting the irradiance uniformity and high efficiency cannot be obtained
information acquisition efficiency. To realize a larger imaging by a simple superposition scheme [24].
area, the main challenge of large-area chlorophyll fluorescence In an illumination design of the system, to improve light
imaging measurement is how to achieve a large-scale imaging energy utilization, the lighting energy should focus on the
area and ensure even illumination over the entire imaging detection area as much as possible. We tackle this issue by
area [20]. We tried to tackle this challenge through the design LED lighting spot reshaping.
of the optical part of the system. Figure 2(a) shows the layout of a single lighting module for
In this study, we introduce a crop population growth lighting spot reshaping. Each module consists of an LED chip,
information detection system based on modulated chlorophyll a light pipe, and a relay lens. The LED chip provides the light
fluorescence detection. The imaging detection area of the sys- energy for chlorophyll fluorescence imaging measurement. The
tem is up to 45 cm × 34 cm. Uniformity of illumination is light pipe is used to make the lighting illumination uniform
above 85%. The system provides different lighting intensity through multiple reflection of the light and transfer the lighting
by a single light source through modulating the duty cycle spot into a rectangular shape, which has the same aspect ratio as
of the control signal. By two-dimensional scanning of the the CCD sensor. The relay lens images the light pipe exit pupil
detection system, measurement of larger area can be realized. to the lighting area.
This system can perform modulated chlorophyll fluorescence Considering chlorophyll fluorescence locates above 660 nm
induction kinetics measurement and chlorophyll fluorescence [7] and exciting light should be filtered easily, a blue LED
parameter imaging. Based on the statistical analysis of fluores- (460 nm center wavelength) is used. To achieve the expected
cence parameter images, the overall crop physiological status high saturating pulse intensity (I S ) within the imaging area (S),
can be obtained. This new developed system, together with the number of LEDs used should satisfy the condition
the corresponding data analysis technique, can significantly
reduce the crop physiological status judge error, improve mon- αN Pσ 460
≥ IS; (1)
itoring efficiency of crop growing condition, and analyze crop K 460 E 460 S
growth condition more comprehensively. The system is also
promising in high throughput application scenarios, such as
plant phenotyping.

2. INSTRUMENT DESIGN
In the progress of modulated chlorophyll fluorescence
detection, the system should provide different lighting intensity
levels [9]. Meanwhile, a large imaging area, uniform illumina-
tion, and high light energy utilization are also critical. The
block diagram of our system is shown in Fig. 1. The system
consists of an optical system, an electronic control system,
and a data acquisition and analysis system. The detailed designs
of the subsystems are described below.
A. Optical System
The optical system is designed to have a large imaging area,
even illumination, and high light energy utilization. We chose
a high-power LED as the lighting source. In traditional illumi-
nation design, the irradiance at a target plane point is formed by
the irradiance sum of all units [21] or the neighboring units Fig. 2. Layout of (a) single lighting module and (b) simulated
[22,23]. As a result, illumination performance with both high irradiance distribution of the single module.
 9764 Vol. 56, No. 35 / December 10 2017 / Applied Optics Research Article

where α is the energy transfer efficiency, N is the number of

LEDs, P is the power of a single LED, σ 460 is the luminous
efficiency of the blue LED with 460 nm luminous center wave-
length, K 460 is the number of lumens equivalent to radiant
energy of 460 nm wavelength, and E 460 is the photon energy
of 460 nm wavelength.
Saturation pulse intensity is above 6000 μmol photons
m−2 s−1 [9]. When energy transfer efficiency is 50% and a
100 W LED is used, the relationship between LED numbers
and illumination area can be obtained. Considering crop and
seedling size and imaging distance, we chose 16 LEDs and
45 cm × 34 cm illumination area. Meanwhile, by using a
CCD sensor with a size of 2/3 in (8.8 mm × 6.6 mm). and
camera lens with standard view, 60 cm lighting distance can
be calculated.
The high-power blue LED used here has a chip-surface-
emitting size of 19 mm × 19 mm. The entrance size of the
light pipe is 22 mm × 29 mm, and its length is 60 mm.
Figure 2(b) shows the simulated irradiance distribution of a
single module. Each lighting module has a 45 cm × 34 cm
illumination area, and the system illumination area should
be obtained through the superposition of 16 modules.
Considering the physical size of the single lighting module
and CCD camera, 16 lighting modules are divided into two
circles around the CCD. Focusing of each module on the
system detection area will cause the incline of each module,
decreasing the illumination uniformity. So, the 16 lighting
modules should be arranged symmetrically to eliminate the non-
uniformity as much as possible. By optimizing the module in-
cline angle and the radius of each circle, the final design of the
optical system structure can be obtained as shown in Figs. 3(a)
and 3(b), where R 1 is 120 mm, R 2 is 200 mm, α1 is 9.8°, and α2
is 16.1°. From the simulation results in Fig. 3(c) we can see the
illumination uniformity within the system lighting area. The Fig. 3. (a) and (b) Layout of lighting module arrangement and
energy outside the 45 cm × 34 cm lighting area is minimized. chlorophyll fluorescence imaging system; (c) simulated irradiance
distribution.
B. Electronics
For modulated the chlorophyll fluorescence measurement tech-
nique, measuring light, actinic light, and saturating light have a with Fig. 4(b)]. Owing to the large power of the LED light
different lighting intensity requirement. In the proposed sys- source, a silicon-controlled rectifier (SCR) is used to switch
tem, different lighting intensity is achieved through adjusting the power of LEDs. When FPGA output (PIN 1) is a low level,
the duty cycle of the lighting pulse sequence. the optoisolator is cut off and SCR conducts. Then the LED
Measuring light is used to excite fluorescence of chlorophyll. array is switched on. And when FPGA output is a high level,
The intensity of measuring light should not be too high to cause the LED array is switched off.
photosynthesis (<1 μmol photons m−2 s−1 ). When the duty The image acquisition signal (PIN 2) is synchronized to the
cycle of the light pulse is close to 1, the lighting intensity satisfies measuring light control pulse. Just like the synchronous detec-
the saturating light requirement (>6000 μmol photons m−2 s−1 ). tion technique, CCD acquires images to coincide only with the
So the pulse width of measuring light is about 30 μs when the fluorescence excited by the measuring light pulse. This ensures
frequency of the measuring light pulse is 5 Hz. For higher image the fluorescence signals will not be interfered with by actinic
acquisition frequency, the pulse width of measuring light should light and saturating light. For the different actinic light inten-
be reduced accordingly. Actinic light aims to drive photosynthe- sity requirement, average light intensity should satisfy the
sis (several hundred μmol photons m−2 s−1 ). It should be requirement by modulating the duty cycle. By inserting the
adjustable to simulate a different lighting environment. measuring light pulse in actinic light and saturating light,
Figure 4(a) shows the schematic diagram of the electronic chlorophyll fluorescence images can be observed in real time.
control system. A field-programmable gate array (FPGA) pulse
timing control circuit is used to output control signals. Upon C. Data Acquisition and Analysis
receiving computer instructions (COM), FPGA outputs the A cooled-type CCD camera PCO1300 (PCO, GER) is used in
light control signal (PIN 1) and image acquisition signal the image acquisition system. It has high sensitivity and low
(PIN 2) as Fig. 4(b) shows [signal level in practice is opposite noise. Binning mode 2 × 2 is used to improve the signal readout
 Research Article Vol. 56, No. 35 / December 10 2017 / Applied Optics 9765

Table 1. Specification of Crop Population Growth

Information Detection System
LED light Wavelength 460 nm
source Total power 1600 W
Detection area 45 cm × 34 cm
Working 60 cm
distance
CCD camera Image 696 × 520 (2 × 2 binning)
resolution
Dynamic range 12 bit
Filter Longpass > 660 nm
Measuring light Intensity 0.1 μmol photons m−2 s−1
Frequency 1–10 Hz
Actinic light Intensity Maximum 2000 μmol photons
m−2 s−1
Saturating light Intensity 6000 μmol photons m−2 s−1
Software 1) Chlorophyll fluorescence parameter image
pseudocolor display
2) Chlorophyll fluorescence kinetics curve real-time
display
Fig. 4. Schematic diagram of (a) electronic control system and 3) Chlorophyll fluorescence parameter distribution
(b) electrical pulse sequence generated by FPGA pulse timing control curve
circuit.

value distribution of the images, the plant area should be

segmented from the images first. Owing to the small pixel val-
speed and signal-to-noise ratio. The final image resolution is ues of the non-plant area, we set a threshold to distinguish the
696 × 520 pixels in a 45 cm × 34 cm area. A longpass filter plant area and the non-plant area. The area whose pixel values
is used to separate the excitation light (460 nm) from the are above the threshold is taken as the plant area. Through the
chlorophyll fluorescence (above 660 nm). Specification of pixel values count of the plant area, the number of different
the system is shown in Table 1. pixel values can be obtained. The parameter distribution curve
Before the beginning of the modulated chlorophyll fluores- can be described as
cence measurement, adequate dark adaptation (>20 min) h  i
should be conducted for the crop. The process of the modulated Sum V − 1n < I ≤ V
chlorophyll fluorescence measurement is as follows. First, mea- PV   ; (2)
Sum0 < I ≤ 1
suring light (about 0.1 μmol photons m−2 s−1 ) is opened, and we
can get the minimal level of fluorescence F o . Second, saturating where P is the proportion of parameter value V , I is the pixel
light (>6000 μmol photons m−2 s−1 ) is opened, photosynthesis gray value of plant area, n is the number of points on the curve,
is inhibited temporarily, and we get the maximal level of fluo- and Sum is the total number of I that conforms to the
rescence F m . Third, actinic light (several hundred μmol photons constraints. Based on Eq. (2) we can finally get the parameter
m−2 s−1 ) is opened to simulate the ambient light condition. distribution curve.
When the photosynthesis is stable, we can get the stable fluores-
cence F and the maximal level of fluorescence after light adap- 3. EXPERIMENTS
tion F m0 . With the above fluorescence parameters, maximal A. System Performance
quantum efficiency of PSII F v ∕F m  F m − F o ∕F m , PSII The photograph of the proposed crop population growth
operating efficiency ΦPSII  F m0 − F ∕F m0 , and other photosyn- Information detection system is shown in Fig. 5. To minimize
thetic parameters can be obtained, an important basis to judge the influence of ambient light, modulated chlorophyll fluores-
the plant physiological status. Meanwhile, chlorophyll fluores- cence measurement is carried out at night. The single detection
cence parameter images are processed with pseudo color to area is 45 cm × 34 cm. Through scanning of the x, y axes
enhance people’s ability to identify the details of the images. driven by the scanning mechanism, detection of larger area
The software supporting modulated chlorophyll fluores- can be achieved. The movement of the system along the z axis
cence measurement allows real-time display of the chlorophyll can make the detection suitable for the crops at different
fluorescence image and induction kinetics curve. For the mea- heights.
surement of the chlorophyll fluorescence induction kinetics Through the measurement of the detection area illuminance
curve, the region of interest (ROI) should be selected first. distribution, we have tested the illuminance uniformity of the
By real-time calculating the mean gray value of the pixels within area. The results are shown in Fig. 6. The illuminance uniform-
the ROI, the kinetics curve will be displayed. ity is defined as the minimum illuminance divided by the aver-
For further analysis of chlorophyll fluorescence parameter age illuminance. So, the illuminance uniformity of the area is
images, the software also supports the fluorescence parameter 85.6%. Meanwhile, we can notice that the low illuminance is
distribution curve to show the change trend of the fluorescence mostly in the corner of the detection area. The illuminance
parameters. Before counting the fluorescence parameter pixel distribution in the center can be more uniform.
 9766 Vol. 56, No. 35 / December 10 2017 / Applied Optics Research Article

Fig. 6. Measurement of the detection area illuminance distribution.

The illumination value measured is normalized to the maximum
illumination.

D. Nitrogen Deficiency Treatment

Fig. 5. Photograph of (a) the crop population growth information Cucumber seedlings with two true leaves were transplanted
detection system and (b) the working state of the system. Modulated into plastic pots filled with perlite till three true leaves. Plants
chlorophyll fluorescence measurement is conducted at night to ensure of control and treatment groups were watered with Hoagland’s
adequate dark adaptation and avoid the interference of sunlight. nutrient solution every two days at first, and then the treatment
plants were watered with nitrogen deficient nutrient solution

B. Plant Material
Seedlings of cucumber (Cucumis sativus L. cv. Jinyan no. 4)
were used. Seeds were germinated in a growth medium filled
with a mixture of peat and vermiculite (3:1, v/v) in trays in
a growth chamber. When the second true leaf was fully
expanded, seedlings were transferred to a nonheated green-
house (Department of Horticulture, Zhejiang University)
for further growth under natural conditions. The average
temperature is 30°C in the daytime, and 23°C at night, and
the relative humidity was 70% to 80%. When it is rainy or
cloudy, the plants were supplemented with an LED source
(200 μmol m−2 s−1 , 12 h/12 h light cycle).
C. Water Deficiency Treatment
Cucumber seedlings with two true leaves were transplanted
into plastic pots (15 cm diameter and 15 cm depth, one seed-
ling per pot) containing the peat and vermiculite (3:1, v/v) till
three true leaves. The stressed plant was stopped in supplying
water to control the growth medium with proper moisture con-
tent, which declines slowly over time. The last time watering Fig. 7. Chlorophyll fluorescence induction kinetics curves of cu-
was designed as on day one. cumbers in drought condition on (a) the first day and (b) the fifth day.
 Research Article Vol. 56, No. 35 / December 10 2017 / Applied Optics 9767

Fig. 8. Chlorophyll fluorescence imaging of cucumbers under nitrogen deficiency. ΦPSII and F v ∕F m showed the decrease trend as nitrogen
deficiency degree deepened. The white rectangular box in the figure shows the area of 10 cm × 10 cm.

(NO−1
3 was replaced by Cl
−1
in Hoagland’s nutrient solution). move to the direction of small numerical values as the degree of
The last time watering normal Hoagland’s nutrient solution nitrogen deficiency progresses and the distribution broadens.
was designed as on day one. The broadening is due to co-existence of both healthy and
diseased tissues. We can understand the nitrogen deficiency de-
E. Chlorophyll Fluorescence Induction Kinetics velopment process quantitatively from the curves. Meanwhile,
Curve
large-area chlorophyll fluorescence images and parameter
With the lighting pulse controlled by the electronic signal, the
system can accomplish the measurement of the chlorophyll
fluorescence induction kinetics curve. From the curve not only
can we calculate chlorophyll fluorescence parameters, we can
also observe the process of fluorescence intensity being stable.
Except for chlorophyll fluorescence parameters calculated
based on the curve, the trend of the curve being stable can also
reflect the physiological status of the plants. From the experi-
ments with chlorophyll fluorescence induction kinetics curve
measurement of cucumbers in drought condition, we can
see how the drought affected the curve (Fig. 7). As the drought
degree deepens from the first day to the fifth day, ΦPSII de-
creases from 0.4 to 0.34. Meanwhile, the time that fluorescence
tends to stabilize becomes longer. The time that fluorescence
intensity declines to 60% has become about 150 s on the fifth
day from about 100 s on the first day. The drought condition
made cucumber photosynthetic efficiency and the ability to
adapt the light environment decrease. Chlorophyll fluorescence
parameters and the change trend of the chlorophyll fluores-
cence induction kinetics curve can help estimate the plant
physiological condition together.
F. Chlorophyll Fluorescence Imaging
The system is useful in applications where we screen a large area
of crops. An illustration of such application is shown in Fig. 8.
Based on chlorophyll fluorescence imaging in a large area, we
can see the development progress of nitrogen deficiency of the
cucumbers. The chlorophyll fluorescence images of ΦPSII
and F v ∕F m showed a decrease as the nitrogen deficiency degree
deepens. Meanwhile, we can observe the heterogeneity of ΦPSII
and F v ∕F m .
From Fig. 8 we can see the abnormal parameter region di-
rectly, and from parameter distribution curves shown in Fig. 9
we can see the change of chlorophyll fluorescence parameters Fig. 9. Parameter distribution curves of cucumbers in nitrogen de-
from another angle. When the cucumber seedlings are in nitro- ficiency. Parameter distribution curves show the change trend of ΦPSII
gen deficiency condition, ΦPSII and F v ∕F m distribution curves and F v ∕F m more directly.
 9768 Vol. 56, No. 35 / December 10 2017 / Applied Optics Research Article

1) The even illumination is designed at a specific distance,

so the measurement should be proceeded on crop canopy. The
leaves locating on different levels of one plant may be under
different illumination intensity. For F v ∕F m, when saturating
light is strong enough, photosynthesis will be inhibited anyway.
Although the fluorescence intensity received is different as the
measuring light intensity differs, relative parameter F v ∕F m will
not be influenced basically. For ΦPSII, various actinic light in-
tensity will cause a different photosynthesis state, but the rela-
tionship is not consistent [25]. As actinic light intensity
increases, the difference of ΦPSII will decrease. So for the leaves
locating on different levels, higher actinic light intensity should
be used to reduce the ΦPSII difference.
2) To minimize the smear effect, which happens when the
CCD has been overexposed [19], especially when a saturating
pulse is used, the delay between saturating pulse and measuring
pulse could be increased. Meanwhile, measuring light pulse
intensity could also be increased to improve the signal-to-
noise ratio.
3) In the process of the CCD capturing images, there is a
response time before the CCD starts to collect the images
after it receives the external trigger signal. The typical response
time is about 15 μs. To ensure all of the fluorescence signal
excited by measuring light is received by the CCD, especially
when the measuring light pulse width is short, a delay time
should be set between measuring light control signal and image
acquisition signal. Alternatively, the exposure time of the
CCD can be set longer than the measuring light pulse width.
These two ways will make the fluorescence signal be acquired
accurately.
4) The pulse width of measuring light can be reduced for
higher frequency of fluorescence signal acquisition. But if the
single measuring pulse is too weak, the dynamic range of the
Fig. 10. Parameter distribution curves of cucumbers within the
CCD cannot be made full use of and the signal-to-noise ratio
white rectangular box in Fig. 8. It is hard to estimate the physiological
of the signal will decrease. The pulse width of measuring
status change trend of the crop from the curves.
light should not be too wide either to ensure the maximal
level of fluorescence F m not exceed the dynamic range of
CCD.
distribution curves can help us obtain the overall physiological In the progress of data acquisition and analysis, we accom-
status of the crop. plish the measurement of the chlorophyll fluorescence induc-
But in the case that the detection area is small, such as tion kinetics curve through the selection of the region of
10 cm × 10 cm illustrated in Fig. 8, the detection results interest. Kinetics curve measurement and real-time display of
will cause judgment error. Figure 10 shows the parameter dis- the whole fluorescence image requires higher processing capacity
tribution curves of cucumbers within the white rectangular of the computer. Based on the kinetics curve, a variety of chloro-
box in Fig. 8. We can hardly obtain the physiological status phyll fluorescence parameters can be calculated that help to
change trend of the crop from the curves in Fig. 10. The analyze the physiology of photosynthesis. Meanwhile, the curve
small sampling detection area will enlarge the fluctuation of change trend also reflects the plant physiological status. The
the crop. A small imaging area is not suitable for crop physio- curve change trend shows the plant’s adjusting ability to light
logical detection. We could reduce the judgment error by the environment.
measurement of multiple samples, but the time cost will In the above application, the system with a 45 cm × 34 cm
increase. detection area can screen and analyze seven or eight cucumber
seedlings in a single measurement, which is beneficial to many
application scenarios, such as high throughput phenotypic
4. DISCUSSION analysis. It is helpful to improve the efficiency, effectiveness,
The system we introduce here can perform modulated chloro- and economy of cultivars in plant breeding.
phyll fluorescence measurement on a large area. We can get Based on large-area imaging detection, the heterogeneity of
more data volume from large-area imaging. This is very helpful photosynthesis between different plants can be shown. This can
for plant physiology diagnosis in greenhouse, field, and other help to locate the abnormal plants or region. Through the
practical circumstances. There are some details about the statistical analysis of chlorophyll fluorescence parameters, we
system we should take care of, as follows: can estimate the crop’s overall physiological status.
 Research Article Vol. 56, No. 35 / December 10 2017 / Applied Optics 9769

5. CONCLUSION 10. A. Stirbet and Govindjee, “On the relation between the Kautsky
effect (chlorophyll A, fluorescence induction) and photosystem II:
In this study, we introduce a new system for crop population basics and applications of the OJIP fluorescence transient,”
growth information detection with chlorophyll fluorescence J. Photochem. Photobiol. B 104, 236–257 (2011).
imaging. This system has characteristics of even illumination, 11. U. Schreiber, U. Schliwa, and W. Bilger, “Continuous recording of
photochemical and non-photochemical chlorophyll fluorescence
sufficient saturating pulse intensity, and large detection area. quenching with a new type of modulation fluorometer,” Photosynth.
This system, together with the corresponding data analysis Res. 10, 51–62 (1986).
technique, can be helpful in reducing the crop physiological 12. L. Nedbal, J. Soukupová, D. Kaftan, J. Whitmarsh, and M. Trtílek,
status judgment error, improving monitoring efficiency of “Kinetic imaging of chlorophyll fluorescence using modulated light,”
the crop growing condition. This system is helpful in studying Photosynth. Res. 66, 3–12 (2000).
13. B. Genty and S. Meyer, “Quantitative mapping of leaf photosynthesis
the heterogeneity of crop photosynthesis activity and is prom- using chlorophyll fluorescence imaging,” Aust. J. Plant Physiol. 22,
ising in high throughput application scenarios. We can use a 277–284 (1995).
chlorophyll fluorescence induction kinetics curve, chlorophyll 14. K. Oxborough and N. R. Baker, “An instrument capable of imaging
fluorescence imaging, and chlorophyll fluorescence parameter chlorophyll a fluorescence from intact leaves at very low irradiance
and at cellular and subcellular levels of organization,” Plant Cell
distribution curves to analyze the plant physiological status.
Environ. 20, 1473–1483 (1997).
We can understand crop growth condition more comprehen- 15. H. K. Lichtenthaler and F. Babani, “Detection of photosynthetic activity
sively based on the analysis. and water stress by imaging the red chlorophyll fluorescence,” Plant
Physiol. Biochem. 38, 889–895 (2000).
Funding. 863 Program (2012AA10A503). 16. A. Porcarcastell, E. Pfündel, J. F. Korhonen, and E. Juurola, “A new
monitoring PAM fluorometer (moni-PAM) to study the short- and long-
term acclimation of photosystem II in field conditions,” Photosyn. Res.
REFERENCES 96, 173–179 (2008).
1. P. Andrade-Sanchez, M. A. Gore, J. T. Heun, K. R. Thorp, A. E. 17. H. M. Kalaji, V. Goltsev, K. Bosa, S. I. Allakhverdiev, R. J. Strasser,
Carmo-Silva, A. N. French, M. E. Salvucci, and J. W. White, and Govindjee, “Experimental in vivo measurements of light emission
“Development and evaluation of a field-based high-throughput pheno- in plants: a perspective dedicated to David Walker,” Photosyn. Res.
typing platform,” Funct. Plant Biol. 41, 68–79 (2014). 114, 69–96 (2012).
2. J. W. White, P. Andradesanchez, M. A. Gore, K. F. Bronson, T. A. 18. U. Schreiber, Pulse-Amplitude-Modulation (PAM) Fluorometry and
Coffelt, M. M. Conley, K. A. Feldmann, A. N. French, J. T. Heun, Saturation Pulse Method: An Overview. Chlorophyll a Fluorescence
D. J. Hunsaker, M. A. Jenks, B. A. Kimball, R. L. Roth, R. J. Strand, (Springer, 2004).
K. R. Thorp, G. W. Wall, and G. Wang, “Field-based phenomics 19. X. Johnson, G. Vandystadt, S. Bujaldon, F. A. Wollman, R. Dubois, P.
for plant genetics research,” Field Crops Res. 133, 101–112 (2012). Roussel, J. Alric, and D. Béal, “A new setup for in vivo fluorescence
3. C. Klukas, D. Chen, and J. M. Pape, “Integrated analysis platform: imaging of photosynthetic activity,” Photosyn. Res. 102, 85–93
an open-source information system for high-throughput plant pheno- (2009).
typing,” Plant Physiol. 165, 506–518 (2014). 20. E. H. Murchie and T. Lawson, “Chlorophyll fluorescence analysis:
4. S. Sankaran, L. R. Khot, C. Z. Espinoza, S. Jarolmasjed, V. R. a guide to good practice and understanding some new applications,”
Sathuvalli, G. J. Vandemark, P. N. Miklas, A. H. Carter, M. O. J. Exp. Bot. 64, 3983–3998 (2013).
Pumphrey, N. R. Knowles, and M. J. Pavek, “Low-altitude, high- 21. I. Moreno, M. Avendañoalejo, and R. I. Tzonchev, “Designing light-
resolution aerial imaging systems for row and field crop phenotyping: emitting diode arrays for uniform near-field irradiance,” Appl. Opt.
a review,” Eur. J. Agron. 70, 112–123 (2015). 45, 2265–2272 (2006).
5. N. R. Baker and E. Rosenqvist, “Applications of chlorophyll fluores- 22. J. W. Whang, Y. Y. Chen, and Y. T. Teng, “Designing uniform illumi-
cence can improve crop production strategies: an examination of nation systems by surface-tailored lens and configurations of led
future possibilities,” J. Exp. Bot. 55, 1607–1621 (2004). arrays,” J. Disp. Technol. 5, 94–103 (2009).
6. R. P. Barbagallo, K. Oxborough, K. E. Pallett, and N. R. Baker, “Rapid, 23. Z. Qin, K. Wang, F. Chen, X. Luo, and S. Liu, “Analysis of condition for
noninvasive screening for perturbations of metabolism and plant uniform lighting generated by array of light emitting diodes with large
growth using chlorophyll fluorescence imaging,” Plant Physiol. 132, view angle,” Opt. Express 18, 17460–17476 (2010).
485–493 (2003). 24. R. Wu, Z. Zheng, H. Li, and X. Liu, “Optimization design of irradiance
7. E. Govindje, “Sixty-three years since Kautsky: chlorophyll a fluores- array for led uniform rectangular illumination,” Appl. Opt. 51,
cence,” Funct. Plant Biol. 22, 131–160 (1995). 2257–2263 (2012).
8. W. L. Butler, “Energy distribution in the photochemical apparatus of 25. U. Rascher, M. Liebig, and U. Lüttge, “Evaluation of instant light-
photosynthesis,” Ann. Rev. Plant Physiol. 29, 345–378 (1978). response curves of chlorophyll fluorescence parameters obtained
9. N. R. Baker, “Chlorophyll fluorescence: a probe of photosynthesis in with a portable chlorophyll fluorometer on site in the field,” Plant
vivo,” Ann. Rev. Plant Biol. 59, 89–113 (2008). Cell Environ. 23, 1397–1405 (2000).