NON-DESTRUCTIVE GROWTH MEASUREMENT
NON-DESTRUCTIVE GROWTH MEASUREMENT
OF SELECTED VEGETABLE SEEDLINGS USING
OF SELECTED VEGETABLE SEEDLINGS USING
MACHINE VISION
MACHINE VISION
Ta-Te Lin, Sheng-Fu Cheng, Tzu-Hsiu Lin, Meng-Ru Tsai
Ta-Te Lin, Sheng-Fu Cheng, Tzu-Hsiu Lin, Meng-Ru Tsai
Department of Agricultural Machinery Engineeri
Department of Agricultural Machinery Engineeri
ng,
ng,
National Taiwan University,
National Taiwan University,
Taipei, Taiwan, ROC
INTRODUCTION
INTRODUCTION
Plant growth measurement and
Plant growth measurement and
modeling
modeling
Machine vision technique
Machine vision technique
Seedling characteristics
Seedling characteristics
OBJECTIVES
OBJECTIVES
Image processing algorithm
Image processing algorithm
development
development
Growth measurements of selected
Growth measurements of selected
vegetable seedlings
vegetable seedlings
Model parameter determination and
Model parameter determination and
simulations
SYSTEM IMPLEMENTATION
SYSTEM IMPLEMENTATION
` Stepping motor Back-lighting apparatus Seedling Rotary stage Tripod TripodDesktop computer CCD camera
Stepping motor driver
SEEDLING CHARACTERISTICS
SEEDLING CHARACTERISTICS
Stem length
Stem length
Height
Height
Span
Span
Total leaf area
Total leaf area
Top fresh weight
Top fresh weight
Top dry weight
Top dry weight
IMAGE PROCESSING ALGORITHM
IMAGE PROCESSING ALGORITHM
Start
Threshold
Skeletonize
Start Tracing
Find root node, set it as the father node, and mark it.
1. Find the child nodes which have not been marked as the father node. 2. Set the pointer to the
father node. 3. Set it as marked.
Is the child node a branch node?
1. Set the type of the child node as branch node. 2. Set itself as the
father node. Read image
Yes
No
Set the type of the child node as termninal node.
RESULT OF NODE TRACING
RESULT OF NODE TRACING
Calibration of cabbage top fresh weight from
Calibration of cabbage top fresh weight from
seedling projection area.
seedling projection area.
Y = -2x10-8X2 +1x10-3X + 0.023 R2 = 0.950 0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 2.25 2.50 0 250 500 750 1000 1250 1500 1750 2000 2250 Projection Area (mm2) T op F re sh W ei gh t ( g) Y = -2x10-8X2 +1x10-3X + 0.023 R2 = 0.950 0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 2.25 2.50 0 250 500 750 1000 1250 1500 1750 2000 2250 Projection Area (mm2) T op F re sh W ei gh t ( g)
Calibration of cabbage top dry weight from
Calibration of cabbage top dry weight from
seedling projection area.
seedling projection area.
Y = 7x10-9 X2 + 7x10-5 X + 2x10-5 R2 = 0.917 0.00 0.05 0.10 0.15 0.20 0.25 0 250 500 750 1000 1250 1500 1750 2000 2250 Projection Area (mm2) T op D ry W ei gh t ( g) Y = 7x10-9 X2 + 7x10-5 X + 2x10-5 R2 = 0.917 0.00 0.05 0.10 0.15 0.20 0.25 0 250 500 750 1000 1250 1500 1750 2000 2250 Projection Area (mm2) T op D ry W ei gh t ( g)
Calibration of cabbage total leaf area from
Calibration of cabbage total leaf area from
seedling projection area.
seedling projection area.
Y = 3x10-4 X2 + 2.633 X - 82.28 R2 = 0.955 0 1000 2000 3000 4000 5000 6000 7000 8000 0 250 500 750 1000 1250 1500 1750 2000 2250 Projection Area (mm2) T ot al L ea f A re a (m m 2 ) Y = 3x10 -4 X2 + 2.633 X - 82.28 R2 = 0.955 0 1000 2000 3000 4000 5000 6000 7000 8000 0 250 500 750 1000 1250 1500 1750 2000 2250 Projection Area (mm2) T ot al L ea f A re a (m m 2 )
Y = 1x10-8 X2 + 8x10-4 X + 3x10-6 R2 = 0.9683 0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00 4.50 5.00 5.50 0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 5500 6000 6500 T op F re sh W ei gh t (g ) Projection Area (mm2) Y = 1x10-8 X2 + 8x10-4 X + 3x10-6 R2 = 0.9683 0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00 4.50 5.00 5.50 0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 5500 6000 6500 T op F re sh W ei gh t (g ) Projection Area (mm2)
Calibration of amaranth top fresh weight from
Calibration of amaranth top fresh weight from
seedling projection area.
Calibration of amaranth top dry weight from
Calibration of amaranth top dry weight from
seedling projection area.
seedling projection area.
Y = 3x10-9 X2 + 7x10-5 X - 0.0113 R2 = 0.9316 0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50 0.55 0.60 0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 5500 6000 6500 Projection Area (mm2) T op D ry W ei gh t (g ) Y = 3x10-9 X2 + 7x10-5 X - 0.0113 R2 = 0.9316 0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50 0.55 0.60 0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 5500 6000 6500 Projection Area (mm2) T op D ry W ei gh t (g )
Calibration of amaranth total leaf area from
Calibration of amaranth total leaf area from
seedling projection area.
seedling projection area.
Y = 6x10-5 X2 + 2.7589 X - 214.46 R2 = 0.9799 0 2000 4000 6000 8000 10000 12000 14000 16000 0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 Projection Area (mm2) T ot al L ea f A re a (m m 2 ) Y = 6x10-5 X2 + 2.7589 X - 214.46 R2 = 0.9799 0 2000 4000 6000 8000 10000 12000 14000 16000 0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000 Projection Area (mm2) T ot al L ea f A re a (m m 2 )
Calibration of kale top fresh weight from
Calibration of kale top fresh weight from
seedling projection area.
seedling projection area.
Y = 8x10-8 X2 + 7x10-4 X - 0.024 R2 = 0.905 0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00 4.50 5.00 5.50 0 500 1000 1500 2000 2500 3000 3500 4000 4500 Projection Area (mm2) T op F re sh W ei gh t (g ) Y = 8x10-8 X2 + 7x10-4 X - 0.024 R2 = 0.905 0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00 4.50 5.00 5.50 0 500 1000 1500 2000 2500 3000 3500 4000 4500 Projection Area (mm2) T op F re sh W ei gh t (g )
Calibration of kale top dry weight from
Calibration of kale top dry weight from
seedling projection area.
seedling projection area.
Y = 2x10-8 X2 + 5x10-5 X + 0.0075 R2 = 0.846 0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50 0.55 0.60 0.65 0 500 1000 1500 2000 2500 3000 3500 4000 4500 Projection Area (mm2) T op D ry W ei gh t (g ) Y = 2x10 -8 X2 + 5x10-5 X + 0.0075 R2 = 0.846 0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50 0.55 0.60 0.65 0 500 1000 1500 2000 2500 3000 3500 4000 4500 Projection Area (mm2) T op D ry W ei gh t (g )
Calibration of kale total leaf area from seedling
Calibration of kale total leaf area from seedling
projection area. projection area. Y = 0.0001 X2 + 1.7988 X - 59.28 R2 = 0.954 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 0 500 1000 1500 2000 2500 3000 3500 4000 Projection Area (mm2) T ot al L ea f A re a (m m 2 )
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2
Manually Measured Top Fresh Weight (g)
T op F re sh W ei gh t f ro m th e Sy st em ( g)
Comparison between manually measured top
Comparison between manually measured top
fresh weight and that determined by the
fresh weight and that determined by the
automatic measurement system.
Comparison between manually measured total
Comparison between manually measured total
leaf area and that determined by the
leaf area and that determined by the
automatic measurement system.
automatic measurement system.
0 1000 2000 3000 4000 5000 6000 0 1000 2000 3000 4000 5000 6000 Manually Measured Total Leaf Area (mm2)
T ot al L ea f A re a fr om th e S ys te m ( m m 2 )
Comparison between manually measured top
Comparison between manually measured top
fresh weight and that determined by the
fresh weight and that determined by the
automatic measurement system.
automatic measurement system.
0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 0.18 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 0.18
Manually Measured Top Dry Weight (g)
T op D ry W ei gh t f ro m th e Sy st em ( g)
Serial images of kale seedlings at various
Serial images of kale seedlings at various
growth stages. (images are not of the same
growth stages. (images are not of the same
scale)
Kale seedlings images from different angles
0.0 0.2 0.4 0.6 0.8 1.0 1.2 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 Time (day) T op F re sh W ei gh t (g )
Top fresh weight of kale seedlings growing
Top fresh weight of kale seedlings growing
under 25/20
under 25/20C. Each curve indicates individual C. Each curve indicates individual
seedling.
Average plant height of kale seedlings grown
Average plant height of kale seedlings grown
under five different day/night temperatures.
under five different day/night temperatures.
0 1 2 3 4 5 6 7 8 9 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 Time (day) A ve ra ge P la nt H ei gh t (c m ) 15 ℃ 20 ℃ 25 ℃ 30 ℃ 35 ℃
Average plant top fresh weight of kale
Average plant top fresh weight of kale
seedlings grown under five different day/night
seedlings grown under five different day/night
temperatures. temperatures. 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 Time (day) A ve ra ge T op F re sh W ei gh t (g ) 15 ℃ 20 ℃ 25 ℃ 30 ℃ 35 ℃
Average top dry weight of kale seedlings
Average top dry weight of kale seedlings
grown under five different day/night
grown under five different day/night
temperatures. temperatures. 0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.10 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 Time (day) A ve ra ge T op D ry W ei gh t (g ) 15 ℃ 20 ℃ 25 ℃ 30 ℃ 35 ℃
Average total leaf area of kale seedlings
Average total leaf area of kale seedlings
growing under five different day/night
growing under five different day/night
temperatures. temperatures. 0 200 400 600 800 1000 1200 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 Time (day) T ot al L ea f A re a (m m 2 ) 15 ℃ 20 ℃ 25 ℃ 30 ℃ 35 ℃
PLANT GROWTH MODELS
PLANT GROWTH MODELS
LOGISTIC MODEL
LOGISTIC MODEL
dY dt Y(1 Y ) t : Time Y : Plant characteristics : Growth constant : Reciprocal of Y when t = Y0 : Y at time = 0
Y = Y
0/ [
Y
0+ ( 1 -
Y
0) e
- t]
PLANT GROWTH MODELS
PLANT GROWTH MODELS
RICHARDS MODEL
RICHARDS MODEL
t : Time
Y : Plant characteristics
: Growth constant
: Reciprocal of Y when t = Y0 : Y at time = 0
: For logistic model, =1
Y = Y
0/ { (
Y
0)
+ [ 1 - (
Y
0)
] e
- t}
1/ dY dt Y Y
[1 (
) ]
Time (day) 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 T op F re sh W ei gh t ( g) 0.0 0.5 1.0 1.5 2.0 2.5 15/12oC 20/15oC 25/20oC 30/25oC 35/30oC Time (day) 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 T op F re sh W ei gh t ( g) 0.0 0.5 1.0 1.5 2.0 2.5 15/12oC 20/15oC 25/20oC 30/25oC 35/30oC
Comparison of regression curves to the
Comparison of regression curves to the
experimental data. Top fresh weight of
experimental data. Top fresh weight of
cabbage seedlings growing under various
cabbage seedlings growing under various
day/night temperatures was used as an
day/night temperatures was used as an
example.
Top fresh weight of cabbage seedlings Temperature (day/night) (g-1) Y0 (g x10-2) (g x10RMSE-2) 15/12ºC 0.159± 0.020 0.601± 0.206 2.80± 0.53 4.30± 1.41 20/15ºC 0.204± 0.064 0.604± 0.335 2.71± 1.57 7.75± 3.34 25/20ºC 0.161± 0.029 0.378± 0.137 4.40± 1.06 9.26± 1.74 30/25ºC 0.172± 0.024 0.551± 0.242 3.61± 1.39 5.25± 2.39 35/30ºC 0.143± 0.026 0.595± 0.386 3.58± 1.08 3.94± 1.32
GROWTH MODEL PARAMETERS
Top fresh weight of cabbage seedlings Temperature (day/night) (g-1) (g x10Y0-2) (g x10RMSE-2) 15/12ºC 0.159± 0.020 0.601± 0.206 2.80± 0.53 4.30± 1.41 20/15ºC 0.204± 0.064 0.604± 0.335 2.71± 1.57 7.75± 3.34 25/20ºC 0.161± 0.029 0.378± 0.137 4.40± 1.06 9.26± 1.74 30/25ºC 0.172± 0.024 0.551± 0.242 3.61± 1.39 5.25± 2.39 35/30ºC 0.143± 0.026 0.595± 0.386 3.58± 1.08 3.94± 1.32
Top fresh weight of kale seedlings Temperature (day/night) (g-1) (g x10Y0-2) (g x10RMSE-2) 15/12ºC 0.162± 0.008 0.912± 0.171 1.45± 0.17 2.34± 0.75 20/15ºC 0.246± 0.053 0.591± 0.318 0.81± 0.45 1.96± 1.24 25/20ºC 0.268± 0.041 0.991± 0.279 0.64± 0.18 2.16± 0.79 30/25ºC 0.217± 0.001 0.853± 0.257 1.15± 0.14 3.57± 0.17 35/30ºC 0.152± 0.004 0.563± 0.082 1.88± 0.44 2.45± 1.54
Top fresh weight of amaranth seedlings Temperature (day/night) (g-1) Y0 (g x10-2) (g x10RMSE-2) 20/15ºC 0.182± 0.012 4.686± 2.076 0.29± 0.07 0.34± 0.25 25/20ºC 0.281± 0.116 1.379± 0.948 0.67± 0.49 4.28± 1.82 30/25ºC 0.374± 0.079 1.280± 0.510 0.29± 0.51 4.61± 3.32 35/30ºC 0.436± 0.050 0.809± 0.108 0.13± 0.06 4.16± 1.35
GROWTH MODEL PARAMETERS
RELATIVE GROWTH RATE, RGR
RELATIVE GROWTH RATE, RGR
LOGISTIC MODEL
LOGISTIC MODEL
RICHARDS MODEL
RICHARDS MODEL
1
1
Y
dY
dt
(
Y
)
1
Y
dY
dt
Y
[
1
(
) ]
Time (day) 2 4 6 8 12 14 16 18 22 24 26 28 0 10 20 30 R el at iv e G ro w th R at e (1 /d ay ) 0.00 0.05 0.10 0.15 0.20 0.25 0.30 15/12oC 20/15oC 25/20oC 30/25oC 35/30oC Time (day) 2 4 6 8 12 14 16 18 22 24 26 28 0 10 20 30 R el at iv e G ro w th R at e (1 /d ay ) 0.00 0.05 0.10 0.15 0.20 0.25 0.30 15/12oC 20/15oC 25/20oC 30/25oC 35/30oC
Predicted relative growth rate of cabbage
Predicted relative growth rate of cabbage
seedling growing under 5 different day/night
seedling growing under 5 different day/night
temperatures using the logistic model.
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 0 5 10 15 20 25 30 Time (day) T op F re sh W ei gh t ( g) Cabbage Amaranth Kale 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 0 5 10 15 20 25 30 Time (day) T op F re sh W ei gh t ( g) Cabbage Amaranth Kale
Comparison of calculated top fresh weight of
Comparison of calculated top fresh weight of
cabbage, amaranth and kale seedlings growing
cabbage, amaranth and kale seedlings growing
at 25/20
0.00 0.05 0.10 0.15 0.20 0.25 0.30 0 5 10 15 20 25 30 Time (day) R el at iv e G ro w th R at e, R G R Cabbage Amaranth Kale 0.00 0.05 0.10 0.15 0.20 0.25 0.30 0 5 10 15 20 25 30 Time (day) R el at iv e G ro w th R at e, R G R Cabbage Amaranth Kale
Comparison of calculated relative growth rate
Comparison of calculated relative growth rate
(RGR) of cabbage, amaranth and kale seedlings
(RGR) of cabbage, amaranth and kale seedlings
growing at 25/20
SEEDLING 3-D RECONSTRUCTION
SEEDLING 3-D RECONSTRUCTION
A B C D
E F G H
SEEDLING 3-D RECONSTRUCTION
SEEDLING 3-D RECONSTRUCTION
CABBAGE SEEDLING
CABBAGE SEEDLING
A B C D
CONCLUSIONS
CONCLUSIONS
A non-destructive machine vision system was A non-destructive machine vision system was
successfully developed for the measurement of
successfully developed for the measurement of
vegetable seedling characteristics. A new algorithm
vegetable seedling characteristics. A new algorithm
for the determination of seedling nodes was
for the determination of seedling nodes was
implemented.
implemented.
3-dimension reconstruction of seedling architecture 3-dimension reconstruction of seedling architecture
can be achieved with the nodal coordinates
can be achieved with the nodal coordinates
determined with the machine vision system.
determined with the machine vision system.
Growth responses of cabbage, kale and amaranth Growth responses of cabbage, kale and amaranth
seedlings under various temperature conditions were
seedlings under various temperature conditions were
measured and compared.
measured and compared.
The dynamic growth responses of selected vegetable The dynamic growth responses of selected vegetable
seedlings were analyzed with logistic and Richards
seedlings were analyzed with logistic and Richards
growth model and the relative growth rates of the
growth model and the relative growth rates of the
seedlings under various conditions were calculated.
FUTURE DEVELOPMENT
FUTURE DEVELOPMENT
Measurement under natural lighting
Measurement under natural lighting
Leaf area index (LAI) determination
Leaf area index (LAI) determination
Extraction of information from serial
Extraction of information from serial
images
images
Modification of the current growth
Modification of the current growth
model
model