Non-destructive growth measurement of selected vegetable seedlings machine version

(1)

NON-DESTRUCTIVE GROWTH MEASUREMENT

OF SELECTED VEGETABLE SEEDLINGS USING

MACHINE VISION

Ta-Te Lin, Sheng-Fu Cheng, Tzu-Hsiu Lin, Meng-Ru Tsai

Department of Agricultural Machinery Engineeri

ng,

National Taiwan University,

Taipei, Taiwan, ROC

(2)

INTRODUCTION



Plant growth measurement and

modeling



Machine vision technique



Seedling characteristics

(3)

OBJECTIVES



Image processing algorithm

development



Growth measurements of selected

vegetable seedlings



Model parameter determination and

simulations

(4)

SYSTEM IMPLEMENTATION

` Stepping motor Back-lighting apparatus Seedling Rotary stage Tripod Tripod

Desktop computer CCD camera

Stepping motor driver

(5)

SEEDLING CHARACTERISTICS



Stem length



Height



Span



Total leaf area



Top fresh weight



Top dry weight

(6)

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.

(7)

RESULT OF NODE TRACING

(8)

RESULT OF NODE TRACING

(9)

Calibration of cabbage top fresh weight from

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)

(10)

Calibration of cabbage top dry weight from

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)

(11)

Calibration of cabbage total leaf area from

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 ）

(12)

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

(13)

Calibration of amaranth top dry weight from

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 )

(14)

Calibration of amaranth total leaf area from

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 )

(15)

Calibration of kale top fresh weight from

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 )

(16)

Calibration of kale top dry weight from

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 )

(17)

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 )

(18)

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

fresh weight and that determined by the

automatic measurement system.

(19)

Comparison between manually measured total

leaf area and that determined by the

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 )

(20)

Comparison between manually measured top

fresh weight and that determined by the

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)

(21)

Serial images of kale seedlings at various

growth stages. (images are not of the same

scale)

(22)

Kale seedlings images from different angles

(23)

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

under 25/20

under 25/20C. Each curve indicates individual _{C. Each curve indicates individual}

seedling.

(24)

Average plant height of kale seedlings grown

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 ℃

(25)

Average plant top fresh weight of kale

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 ℃

(26)

Average top dry weight of kale seedlings

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 ℃

(27)

Average total leaf area of kale seedlings

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 ℃

(28)

PLANT GROWTH MODELS



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

]

(29)

PLANT GROWTH MODELS



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 (



) ]



(30)

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

experimental data. Top fresh weight of

cabbage seedlings growing under various

day/night temperatures was used as an

example.

(31)

Top fresh weight of cabbage seedlings Temperature (day/night)  (g-1₎ Y0 (g x10-2₎ _{(g x10}RMSE-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

(32)

Top fresh weight of cabbage seedlings Temperature (day/night)  (g-1₎ _{(g x10}Y0-2₎ _{(g x10}RMSE-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 x10}Y0-2₎ _{(g x10}RMSE-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 x10}RMSE-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

(33)

RELATIVE GROWTH RATE, RGR



LOGISTIC MODEL



RICHARDS MODEL

1

1 Y

dY

dt





(





Y

)

1 Y

dY

dt

Y





[

1 

(



) ]



(34)

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

seedling growing under 5 different day/night

temperatures using the logistic model.

(35)

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

cabbage, amaranth and kale seedlings growing

at 25/20

(36)

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

(RGR) of cabbage, amaranth and kale seedlings

growing at 25/20

(37)

SEEDLING 3-D RECONSTRUCTION

A B C D

E F G H

(38)

SEEDLING 3-D RECONSTRUCTION



CABBAGE SEEDLING

A B C D

(39)

CONCLUSIONS

 A non-destructive machine vision system was A non-destructive machine vision system was

successfully developed for the measurement of

vegetable seedling characteristics. A new algorithm

for the determination of seedling nodes was

implemented.

 3-dimension reconstruction of seedling architecture 3-dimension reconstruction of seedling architecture

can be achieved with the nodal coordinates

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

measured and compared.

 The dynamic growth responses of selected vegetable The dynamic growth responses of selected vegetable