Earth Observation systems for operational monitoring of crop conditions · 2020. 5. 25. · 1957 - Sputnik spacecraft ... 400 460 520 580 640 700 760 820 880 940 1000 1060 1120 1180

From -omics to phenotyping for crop improvement

Tuesday 26th-Thursday 28th, June 2018

Aula Maggiore - Via Celoria, 2 (MI)

Earth Observation systems for

operational monitoring of crop conditions

Mirco Boschetti CNR-IREA (MI)

Wednesday 27th, June 2018

Intro

Needs of efficient sensing of plants traits

From experimental to real farming condition

Imaging and remote sensing

What is the remote sensing

What can provide for agriculture monitoring

Plant spectral response

Example of application

Mapping, Monitoring and Modelling info for science and management

Spetroradiometric

UAV

satellite

Towards operational downstream service

2

PRESENTATION LAYOUT

Institute for Electromagnetic Sensing of the Environment of CNR develops methodologies and technologies for acquisition, processing, fusion and interpretation of images and data obtained by electromagnetic sensors - operating on satellite, aircraft and in situ

NRM_LAB is a multidisciplinary team working on environmental and agricultural monitoring issues

We study and develop solutions and methods to generate and provide end-user value-added information generated by the acquisition, processing and integration of multisource data

In Copernicus we are dedicated to the research for the creation of "Downstream services" prototypes, especially for the agricultural sector

An operational workflow to assess rice nutritional status based

on satellite remote sensing and smart apps

Nutini et al. 2018 (2018)Computers and Electronics in Agriculture (in press)

Downstream Services for Rice Crop Monitoring in Europe: From Regional to Local Scale Busetto et al. (2017)IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing

Early season weed mapping in rice crops using multi-spectral UAV data

Stroppiana et al. 2018 (2018)nternational Journal of Remote Sensing

ERM

ES m

on

ito

rin

gsy

stem

s

WHO ARE WE: IREA NATURAL RESOURCE MONITORING LABORATORY

INTRO:NEEDS OF EFFICIENT SENSING OF PLANTS RESPONSE TO ENVIRONMENT (G X E)

WHY REMOTE SENSING: ACQUISITION OF NON DESTRUCTIVE QUANTITATIVE INFO

Plant genomic technologies are already well developed a lack of access to plant phenotyping capabilities limits our ability to dissect the genetics of quantitative traits.

Current assessments of phenotype characteristics for disease resistance or stress in breeding programs rely largely on visual scoring by experts, which is time-consuming and can generate bias between different experts and experimental repeats.

High-throughput phenotyping platforms have recently been developed to solve this problem using variety of imaging/sensing methodologies to collect data for quantitative studies (growth, yield and adaptation to biotic or abiotic stress - disease, insects, drought and salinity).

5Lei Li, Qin Zhang and Danfeng Huang (2014) A Review of Imaging Techniques for Plant Phenotyping Sensors 2014, 14, 20078-20111; doi:10.3390/s141120078

IMAGING PLANTS IS MORE THAN JUST ‘TAKING PICTURES’.

The aim of imaging is to measure a phenotype quantitatively through the interaction between light and plants such as reflected photons, absorbed photons, or transmitted photons.

Each component of plant cells and tissues has wavelength-specific absorbance, reflectance, and transmittance properties.

chlorophyll absorbs photons primarily in the blue and red spectral region

water has its primary absorption features in the near and short wavelengths

cellulose absorbs photons in a broad region between 2200 and 2500 nm.

6Noah Fahlgren, Malia Agehan, Ivan Baxter (2015) Lights, camera, action: high-throughput plant phenotyping is ready for a close-up. Current Opinion in Plant Biology 2015, 24:93–99

FROM G X E TO G X E X M

Furthermore, the environmental plasticity of plants creates an interesting interdisciplinary modeling challenge for plant scientists, computer scientists, and statisticians, and resulting models have the potential to guide adjustments in agricultural practices to optimize yield prior to genetic improvements. G X E X M

7Lei Li, Qin Zhang and Danfeng Huang (2014) A Review of Imaging Techniques for Plant Phenotyping Sensors 2014, 14, 20078-20111; doi:10.3390/s141120078

Such monitoring tools are fundamental for breeders and to support phenotyping studies G x E

BASIS OF REMOTE SENSING AND EARTH OBSERVATION

REMOTE SENSING CATEGORICAL & QUANTITATIVE INFORMATION

CROP MONITORING QUESTIONS

Estimates a variable

categorical: which are the crops in the area?

quantitative: how much is the LAI ?

Assess the spatial variation

How much the LAI is varying in the field?

Where are the hot spot anomalous areas?

Quantify the temporal variation

Which is the temporal trend of LAI?

RS: an instrument for research and land management

REMOTE SENSING

RS is a science of obtaining information about an object, area, or phenomenon through the analysis of data acquired by a device that is not in contact with the object, area, or phenomenon under investigation

RS is the non-contact recording of information from the UV, VIS, NIR, MW of the EM spectrum by means of instruments such as cameras, scanners, lasers, linear arrays, and/or area arrays located on platforms such as aircraft or spacecraft, and the analysis of acquired information by means of visual and digital image processing

Energy and material interaction

REMOTE SENSING PRINCIPLE

RS measure measure electromagnetic energy reaching the sensor after the interaction with a surface

Source

Pixel fligth heigth(sensorcharacteristcs)

Different surfaces provide peculiarresponce with the EM reflecting, absorbing and emitting in a different way at differentwavelength

ESA Living Planet Symposium, Prague 2016

REMOTE SENSING: PASSIVE AND ACTIVE SENSORS

ESA Living Planet Symposium, Prague 2016

1840 - Hot air balloon with camera

1909 - Pigeon, light cameras (70 g)

1943 - German missiles V2

1957 - Sputnik spacecraft

1960 - First meteorological satellites

1972 - First Earth-sensing satellite (Landsat)

1980 - Specialized Sensors: Coastal Zone Color Scanner (CZCS), Heat

Capacity Mapping Mission (HCMM), and Advanced Very High Resolution

Radiometers (AVHRR)

1999 - Launch of Ikonos, the first high-resolution commercial satellite

REMOTE SENSING: HISTORY

DRONE AGE

INTERNATIONAL FRAMEWORK

Many opportunities exist for territorial monitoring with Earth Observationsystems in terms of available sensors and international interventions tocoordinate spatial policies

Group on Earth Observations (GEO) coordina la costituzione del Global Earth Observation System of Systems (GEOSS).

Two SAR sensors, 20m, every6 days

Two optical sensors, 10/20m, every 5 days

Optical sensor, 300m, every day

Operational and free of charge

EUROPEAN AND ITALIAN CONTEXT

REMOTE SENSING MEASUREMENTS:

PLANT SPECTRAL RESPONSE

SPECTRAL RESPONSE OF SURFACES

It is possible to discriminate in an image a large number of elements (soil,

vegetation, water, etc.) and to recognize their characteristics (humidity,

state of health, nutrient concentration, etc.) by analyzing the different spectral behavior in the various lengths of wave or their spectral signature

VEGETATION SPECTRAL RESPONSE

Spectral behavior of vegetation depends mainly on two

factors:

the chemical / physical characteristics of the leaves and

other components of the plant

• Chlorophyll content

• Cellular structure

• Water content

the aggregation of the individual elements (leaves,

branches) and the overall structure of the plant (canopy)

• Degree of coverage

• Amount of green biomass

• Architecture of the foliage

• Presence and type of background (soil and weeds)

• phenology

• Health state

• External factors (morphology, source-object-sensor geometry, atmosphere ...)

M.A

. G

om

ara

sca

, 1998

•foliar pigments absorb in the wavelengths of blue and red and reflect in those of the green and it is precisely for this reason that our eyes identify the vegetation of

this color

•structure of the vegetation instead implies that this reflects highly in the near infrared, determining a typical platò in the signature


•graph shows the trend of spectral signatures of vegetation as the chlorophyll content changes

Riflettanza in funzione di Cab

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

0.45

0.5

40

0

47

0

54

0

61

0

68

0

75

0

82

0

89

0

96

0

10

30

11

00

11

70

12

40

13

10

13

80

14

50

15

20

15

90

16

60

17

30

18

00

18

70

19

40

20

10

20

80

21

50

22

20

22

90

23

60

24

30

25

00

l [nm]

r

15

35

55

N 1.5

Cab VARIABLE

Cw 0.015

Cm 0.01

M.

Me

ron

i, 2

001 L

ezi

on

i Co

rso

di T

ele

rile

va

me

nto

Sc

ien

ze A

mb

ien

tali

Un

imib

Leaf chemestry

The signatures are the results of the model (PROSPECT) of the leaf's behavior


•The graph shows the trend of the spectral signatures of vegetation as the water content changes

Riflettanza in funzione di Cw

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

0.45

0.5

40

0

47

5

55

0

62

5

70

0

77

5

85

0

92

5

10

00

10

75

11

50

12

25

13

00

13

75

14

50

15

25

16

00

16

75

17

50

18

25

19

00

19

75

20

50

21

25

22

00

22

75

23

50

24

25

25

00

l [nm]

r

0.005

0.015

0.025

N 1.5

Cab 35

Cw VARIABLE

Cm 0.01

M.

Me

ron

i, 2

001 L

ezi

on

i Co

rso

di T

ele

rile

va

me

nto

Sc

ien

ze A

mb

ien

tali

Un

imib

The signatures are the results of the model (PROSPECT) of the leaf's behavior

Leaf chemestry


Riflettanza in funzione di LAI

0

0.1

0.2

0.3

0.4

0.5

0.6

400

460

520

580

640

700

760

820

880

940

1000

1060

1120

1180

1240

1300

1360

1420

1480

1540

1600

1660

1720

1780

1840

1900

1960

2020

2080

2140

2200

2260

2320

2380

2440

2500

l [nm]

3

5

7

0 isat

1 ihot

40 Theta_s,

40 tl

VARIABLE l

1.5 vai

35 cab

0.015 cw

0.01 cm

0.25 sl

100 vis

1 alph_soil

1 na

0 0

M.

Me

ron

i, 2

001 L

ezi

on

i Co

rso

di T

ele

rile

va

me

nto

Sc

ien

ze A

mb

ien

tali

Un

imib


The signatures are the results of the model (SAIL) of the Canopy behavior

•graph shows the trend of vegetationspectral signatures of as a function of LAI

Plant structure

•graph shows the trend of vegetationspectral signatures of as a function of leaf angles

Plant structure

l

Riflettanza in funzione di MTA

[nm]

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

400

460

520

580

640

700

760

820

880

940

1000

1060

1120

1180

1240

1300

1360

1420

1480

1540

1600

1660

1720

1780

1840

1900

1960

2020

2080

2140

2200

2260

2320

2380

2440

2500

5

40

90

0 isat

1 ihot

40 Theta_s,

VARIABLE tl

5 l

1.5 vai

35 cab

0.015 cw

0.01 cm

0.25 sl

100 vis

1 alph_soil

1 na

0 0

The signatures are the results of the model (SAIL) of the Canopy behavior M.

Me

ron

i, 2

001 L

ezi

on

i Co

rso

di T

ele

rile

va

me

nto

Sc

ien

ze A

mb

ien

tali

Un

imib


VEGETATION SPECTRAL RESPONSE: TEMPORAL DYNAMICS

From the study of the spectral behavior of vegetation a series of

quantitative relationships have been defined between remote sensing

data and vegetation parameters through indices based on the

relationship between the typical absorption and reflection bands

These algebraic relationships are referred to as vegetation indices (VI)

and are based above all on the red and near IR wavelengths (wide or

narrow band)

The VIs are related to the amount of plant biomass, LAI, chlorophyll

concentration, water content etc. and give indications on the state of

health, on crop productivity, on density and coverage and on nutritional

status etc.

VEGETATION INDICES

VEGETATION INDICES

VEGETATION INDEX: BAND COMBINATION

NDVI (Normalized Difference Vegetation Index)

A good index has to emphasise the property under investigation minimising the influence from other factors

0.00

0.10

0.20

0.30

0.40

0.50

400 500 600 700 800 900 1000

Wavelength

Ref

lect

ance

*.0.2

*.1.2

*.2.2

RED

NIR

VEGETATION INDEX: NDVI CONCEPT

4.42

4.19

4.24

4.25

4.34

4.64

4.64

4.34

4.35

5.13

5.18

5.88

1495180 1495200 1495220 1495240 1495260 1495280 1495300 1495320

5018960

5018980

5019000

5019020

5019040

5019060

5019080

1.5

1.9

2.3

2.7

3.1

3.5

3.9

4.3

4.7

5.1

5.5

5.9

6.3

LAI maps from field measurements (1° of July)

VI maps from MIVIS sensor (2° of July)

VEGETATION INDEX: NDVI MAPS VS LAI

NDVI = 0.15Ln(x) + 0.69

R2 = 0.830.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

0 1 2 3 4 5 6 7 8

LAI

ND

VI

a

WDRVI = 0.29Ln(x) + 0.13

R2 = 0.85

-0.5

-0.3

-0.1

0.1

0.3

0.5

0.7

0.9

0 1 2 3 4 5 6 7 8

LAI

WD

RV

I

b

y = 0.15Ln(x) + 0.44

R2 = 0.80

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

0 1 2 3 4 5 6 7 8

LAI

EV

I

c

y = 5.72Ln(x) + 9.64

R2 = 0.65

0

5

10

15

20

25

30

35

40

45

0 1 2 3 4 5 6 7 8

LAI

SR

d

y = 0.07Ln(x) + 0.21

R2 = 0.66

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

0 1 2 3 4 5 6 7 8

LAI

PV

I

e

y = 0.08Ln(x) + 0.27

R2 = 0.710.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

0 1 2 3 4 5 6 7 8

LAI

WD

VI

f

REMOTE SENSING OF AGRICULTURE

By the late 1960s, the first unmanned satellite specifically dedicated to multispectral remote sensing entered the planning stages. NASA carefully designed and constructed, then launched on July 23, 1972 ERTS-1 (Earth Resources Technology Satellite) later renamed Landsat 1.

Since 1974 the LACIE (Large Area

Crop Inventory Experiment) project aimed to provide information on agri-production (USSR)

EARTH OBSERVATION STARTED WITH AGRO-MONITORING

The first satellite image of Lombardy acquired by Landsat 1

the 14th of August 1972, 22 days after the launch

CNR-IREA archive

The first satellite sentinel 2 satellite image acquired the 29th of

June 2015 on the Po Valley, Italy, 6 days after the launch

Copernicus data (2015)/ESA

EARTH OBSERVATION: CONTRIBUTION TO CROP MONITORING

dimensione spaziale dei fenomeni

Visione multispettrale dell’oggetto indagato

Analisi e un monitoraggio multitemporale

Visione sinottica dall’alto del territorio

1000 m

250 m

30 m

> 5 m

< 0.5 m

EARTH OBSERVATION: CONTRIBUTION TO CROP MONITORING

Tipologia delle colture

Date di semina e lavorazioni

23 May

5 June

Variabilità dei suoli

Stato della coltura (vigore/anomalie)

4.4

2

4.1

9

4.2

4

4.2

5

4.3

4

4.6

4 4.6

4

4.3

4

4.3

5

5.1

3

5.1

8

5.8

8

149

51

80

149

520

01

495

22

014

95

24

014

952

60

149

52

80

149

53

00

1495

32

0

50

18

96

0

50

18

98

0

50

19

00

0

50

19

02

0

50

19

04

0

50

19

06

0

50

19

08

0

1.5

1.9

2.3

2.7

3.1

3.5

3.9

4.3

4.7

5.1

5.5

5.9

6.3

7 July

Multi-temporale Multi-sensore

Stato della coltura (stress idrico)

T °C7 July

Sviluppo della coltura

APPLICATION DOMAIN

REMOTELY SENSED CROP PARAMETERS

28/08/2018 37

• Leaf Pigments

• Nitrogen content

• Canopy structural parameters

• Phenology

• Anomaly detection

FIELD: SPECTRO RADIOMETRIC MEASUREMENTS

FE

12

11

10

9

8

7

6

5

4

3

2

1

0 kg/ha

90 kg/ha

180 kg/ha

270 kg/ha

O

N

M

L

M

N

L

O

M

L

O

N

Chl a+b Fc FS-canopy FS-probe

Fila : 12-24-36-60Metr i: 5-10-15

Fila : 18-36-54M e t r i : 1 0

Fila : 18-36-44Met ri : 5 -15

Fila : 36Metri: 10

AISA RGB-FC

A B C D E F

Disegno speriementale

LEAF AND CANOPY MEASUREMENTS: CLOROPHYLL DETECTION

Experimental design: sugar beet

Variabilità controllata generata a seguito di fertilizzazioni differenziate:

4 livelli N (0-90-180-270 kg/ha) randomizzati

2 livelli irrigui (blocchi E, F)

3 repliche

tot 24 parcelle 0.06 ha

Densità semina bietola: interfile di 45 cm, 60 file per parcella

LEAF MEASUREMENTS: CLOROPHYLL DETECTION

Sugar beet (Measuremements acquisition 27/05/05)

ASD_ FS-PRO (Leaf)

Fila : 36

Metri: 10 Tot. 24

Campionamenti di rondellefogliari di 18mm diametro in corrispondenza delle misureper estrazione analitica di Chla+b

Conctact probe accoppiato a FS3 misure spettrali per ogni foglia

LEAF MEASUREMENTS: CLOROPHYLL DETECTION

Leaf spectral responce

Examples of spectral signatures acquired with the contact probe on plants of different levels of fertilization.Effects in the visible region.

0.00

0.10

0.20

0.30

0.40

0.50

0.60

350 850 1350 1850 2350

[nm]

Rifle

ttan

za

0.00

0.10

0.20

0.30

450 550 650 750

N 0%

N 50%

N 100%

N 150%

REP

LEAF MEASUREMENTS: RESULTS

Leaf VIs vs Chl a+b

y = 0.00x + 0.05

R2 = 0.69

0.00

0.02

0.04

0.06

0.08

0.10

0.12

0.14

0.16

0.18

0.20

20 30 40 50chl a+b (mg/cm2)

VI

NDI

y = 1.49e-0.04x

R2 = 0.69

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

0.40

0.45

0.50

20 30 40 50chl a+b (mg/cm2)

VI

MCARI

y = 0.60x + 692.68

R2 = 0.74

706.00

708.00

710.00

712.00

714.00

716.00

718.00

720.00

20 30 40 50chl a+b (mg/cm2)

VI

REP

y = 0.63x + 684.30

R2 = 0.64

700.00

702.00

704.00

706.00

708.00

710.00

712.00

714.00

20 30 40 50chl a+b (mg/cm2)

VI

REP_LAGR

SR NDVI R1 R2 R3 TVI

Chl a+b 0 0 0.49 0.55 0.63 0.21

MTVI2 MCARI TCARI REIP_lagr REIP_lin -

Chl a+b 0.32 0.61 0.64 0.68 0.74 -

Leaf

ns ns

y = 0.06x + 1.39

R2 = 0.62

0.00

0.50

1.00

1.50

2.00

2.50

3.00

3.50

4.00

4.50

20 30 40 50chl a+b (mg/cm2)

VI

750/700

y = -0.18x + 36.09

R2 = 0.20

0.00

5.00

10.00

15.00

20.00

25.00

30.00

35.00

20 30 40 50chl a+b (mg/cm2)

VI

TVI

CANOPY LEVEL: DATA ACQUISITION

Misure radiometriche: 27/05/05 (sorvolo AISA)

ASD-FS-PRO (Canopy)

Fila : 18-36-44

Metri: 10 Tot. 144

0.00

0.10

0.20

0.30

0.40

0.50

0.60

0.70

350 450 550 650 750 850 950

[nm]

Rif

lett

an

za

FO5

FN1

FM4

FL3

CANOPY LEVEL: DATA ACQUISITION

Canopy VIs vs Chl-a+b

Canopy

SR NDVI R1 R2 R3 TVI MTVI2

Chl a+b 0.62 0.69 0.56 0.65 0.66 0.65 0.67

FC 0.80 0.84 0.6 0.70 0.70 0.62 0.63

MTVI2 MCARI TCARI REIP_lin OSAVI TSAVI TCARI/OSAVI

Chl a+b 0.67 0.60 0.6 0.64 0.66 0.64 0.35

FC 0.63 0.46 0.46 0.88 0.72 0.73 0.26

CHL FC

FN1 32.93 64.62

FL3 34.94 83.64

FM4 32.41 54.56

FO5 27.02 35.99

•R2 indici di vegetazione

Examples of spectral signatures on plants of different levels of fertilization.Major effects in the REP and NIR region.Significant contribution of the soil

0%

10%

20%

30%

40%

50%

60%

400 500 600 700 800 900

Lunghezza d'onda(nm)

Rif

lett

an

za

(%

)

N-0

N-150

MCARI-0

MCARI-150

REP-0

REP-150

Chl 28 mg/cm2

Chl 44 mg/cm2

0%

10%

20%

30%

40%

50%

60%

400 500 600 700 800 900

Lunghezza d'onda(nm)

Rif

lett

an

za

(%

) N-0

N-150

MCARI-0

MCARI-150

REP-0

REP-150Chl 28 mg/cm2

Chl 41 mg/cm2

Fc = 28%

Fc = 78%

MCARI > MCARI MCARI < MCARI

LEAF VS CANOPY LEVEL: COMPLEXITY OF DATA ACQUISITION

CANOPY MEASUREMENTS: NITROGEN ESTIMATION

46

Experiment:

• Two vaieties (Gladio, Volano)

• Tree fertilization level two time of

application (0, 80, 160 kg ha-1)

Agronomic parameters

• LAI (transect LAI2000)

• AGB (samples on 20 plants)

• PNC (samples on 6 plants)

0.0

0.1

0.2

0.3

0.4

0.5

350 450 550 650 750 850

l [nm]

Refl

ec

tanc

eJun 16

Jun 24

Jun 30

Jul 07

Jul 21

Jul 26

Spectral

measumrements•FieldSpec FR Pro, (ASD)

• 1nm, 4nm, 350-2500 nm

• Five acquisitions per plot

Experiments was aimed to generate plant growing condition with divergent PNC

and Biomass trend to be investigated by spectral measurementsFertlization: N0

0,0

0,1

0,2

0,3

0,4

0,5

0,6

0,7

0,8

400 500 600 700 800 900 1000 1100 1200

Wavelength [nm]

Re

fle

cta

nc

e

Jun 16

Jun 24

Jun 29

Jul 08

Jul 23

Jul 27

a)

Fertlization: N2

0,0

0,1

0,2

0,3

0,4

0,5

0,6

0,7

0,8

400 500 600 700 800 900 1000 1100 1200

Wavelength [nm]

Re

fle

cta

nc

e

b)

date*Tesi ; LS Means

Current effect: F(5, 35)=8.4386, p=.00003

Effective hypothesis decomposition

Vertical bars denote 0.95 confidence intervals

Tesi

0

Tesi

2

6/17/04 6/25/04 7/1/04 7/9/04 7/22/04 7/28/04

date

-2

-1

0

1

2

3

4

5

6

7

8

9

Bio

[t/

ha

]

***

***

date* Tesi; LS Means

Current effect: F(5, 35)=2.3444, p=.06163

Effective hypothesis decomposition

Vertical bars denote 0.95 confidence intervals

Tesi

0

Tesi

2

6/17/04 6/25/04 7/1/04 7/9/04 7/22/04 7/28/04

date

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

N (

%)

*** ***

**

1° fert 22 jun 2° fert 20 Jul.


NDI (Normalises Difference Index) iperspectral for N estimates

(r2=0.0) (r2

=1.0)

a) b) c)

(r2=0.0) (r2

=1.0)

a) b) c)

PNC: Optimal combinations fall in the VIS (l<700 nm) in the blue-green region where the

photosynthetic pigments have a strong influence. Best NDI l2= 503, l1= 483 (R2=0.65,

***p<0.001)

AGB: High correlation in the VISNIR. Best NDI l1~800nm and l2~600nm. (R2> 0.7;)

λ2=

503 n

m

λ1=483 nm

Band

selection

PNC AGB


NDI vs other indices proposed in literature

y = -0.27ln(x) + 0.903R² = 0.219

0.000

0.200

0.400

0.600

0.800

1.000

1.200

0 0.5 1 1.5 2 2.5 3 3.5

ND

VI(

80

0/6

70

)

PNC %

T0

T2

y = -14.0ln(x) + 726.4R² = 0.165

670.000

680.000

690.000

700.000

710.000

720.000

730.000

740.000

0 0.5 1 1.5 2 2.5 3 3.5

lR

EP

PNC %

T0

T2

y = -1.32ln(x) + 3.592R² = 0.265

0.000

0.500

1.000

1.500

2.000

2.500

3.000

3.500

4.000

4.500

5.000

0 0.5 1 1.5 2 2.5 3 3.5

R1

(70

0/6

70

)

PNC %

T0

T2

y = -0.06ln(x) + 0.146R² = 0.330

0.000

0.050

0.100

0.150

0.200

0.250

0 0.5 1 1.5 2 2.5 3 3.5

MC

AR

I

PNC %

T0

T2

y = -4.69ln(x) + 15.89R² = 0.112

0.000

5.000

10.000

15.000

20.000

25.000

0 0.5 1 1.5 2 2.5 3 3.5

TVI

PNC %

T0

T2

y = -0.05ln(x) + 0.149R² = 0.65

0.000

0.020

0.040

0.060

0.080

0.100

0.120

0.140

0.160

0.180

0.200

0 0.5 1 1.5 2 2.5 3 3.5

ND

I

PNC %

T0

T2



28/08/2018 50

FROM SPECTRA TO IMAGES

AERIAL: UAV MAPPING

Satellite UAVSatellite UAVSatellite UAV

Very high

spatial resolution

Multi-spectralinformation

Satellite UAVSatellite UAV

Very high

spatial resolution

Mission flexibility


Large area coverage

Satellite UAVSatellite UAV

Very high

spatial resolution

Mission flexibility

Less atmospheric and cloud contamination

Time data acquisitionto delivery


Large area coverage

Longer history

Data archives

Cost

cost vs spatial resolution

SATELLITE VS. UAV

Measurements• Crop Rice volano• N -> Destructive, Dualex, PocketN• LAI -> PocketLAI• Reflectance -> Spectroradiometer SR3500

(15/07/14)• UAV acquisition (September 24th, 2014) DJI

S1000 Octocopter & Tetracam ADC Micro

UAV – CROP DEVELOPMENT VS N APPLICATION (RICE)

IDParcella

Kg N/ha04/06/14

Kg N/ha15/07/14

N1.1 20 0

N1.2 20 0

N1.3 20 0

N1.4 20 0

N1.5 20 20

N1.6 20 20

N1.7 20 40

N1.8 20 40

IDParcella

Kg N/ha04/06/14

Kg N/ha15/07/14

N2.1 40 0

N2.2 40 0

N2.3 40 0

N2.4 40 0

N2.5 40 20

N2.6 40 20

N2.7 40 40

N2.8 40 40

IDParcella

Kg N/ha04/06/14

Kg N/ha15/07/14

N3.1 60 0

N3.2 60 0

N3.3 60 0

N3.4 60 0

N3.5 60 20

N3.6 60 20

N3.7 60 40

N3.8 60 40

UAV – CROP DEVELOPMENT VS N APPLICATION (RICE)

28/08/2018 55

Rice

UAV – WEED DETECTION (RICE)

DJI S1000 Octocopter Digital Camera Canon S100 Tetracam ADC Micro

Mutispectral data DSM

UAV – WEED DETECTION (RICE)

28/08/2018 57

Weed proportion Fractional Cover Canopy heigth

UAV: DAMAGE ASSESSMENT (RICE)

58

0 10m

0 10m

0 10m

Riso Centauro

Riso CL26

data 10 agosto 2017

Two experimental fields inVillarasca (PV). (Acqua e Sole S.r.l)

LEGENDA

I RISO CL 26

J RISO CENTAURO

A ACCESTIMENTO

L LEVATA

F FIORITURA

CORRIDOIO

C1, C2, C3, C4 DEFOGLIAZIONE 0%

100 1, 100 2 DEFOGLIAZIONE 100%

50 1, 50 2 DEFOGLIAZIONE 50 %

UAV: DAMAGE ASSESSMENT

VI calculation and damage assessment

59

testimone

Damage

@ steam elongationDamage @Flowering control

sorgente MSAVI RBIAS

0 10m 0 10m 0 10m

Significant relations

between processing of

remote sensing data

and SAPR (ANOVA):

• % defoliation

• Variety

• Phenological Phase

0% 50% 100%

SATELLITE: HIGH RESOLUTION MAPPING

SATELLITE DATA: WITHIN FIELD ANOMALY

28/08/2018 61

Blast disease

Fertilisation plot

Nutritional deficiency

Seed density test

SATELLITE DATA: YIELD VARIABILITY ASSESSMENT

Analysis of satellite images acquired in key phenological phases revelaed significant relation with final yield data measured by a harvester combiner

62

ResaN

dR

E

Resa

Nd

RE

Yield maps

16 Luglio 2014 17 Agosto 2014

VI maps

Ottobre 2014

LAI sampling and foliar nitrogen 'driven' by the 1st satellite image acquiresd

28th June 2015

10/09/2015 Open Day Ente Risi

SATELITE DATA TO SUPPORT SMART SCOUTING

1° July 2015

Biomassindicator

RGB 1 RGB 2 Colori reali

Nitrogen contentindicator

NDVI (Normalised Different Vegetation Index)

CVI (Chlorophyll Vegetation Index)

SATELLITE DATA TO NITROGEN NUTRIZION INDEX (NNI )

NITROGEN NUTRITIONAL INDEX

An operational workflow to assess rice nutritional status based on satellite imagery and smartphone apps

Francesco Nutini a*, Roberto Confalonieri b, Alberto Crema a, Ermes Movedi b, Livia Palearib, Dimitris Stavrakoudis c and Mirco Boschetti a*

Smart Scouting EO data analysis Maps generation

<= 0.7 0.7 - 0.9 0.9 – 1.1 ESU1.1 – 1.3 >= 1.3Cluster a Cluster b Cluster c ESU

PocketLAI

PocketN

Ncrit=

Dilution curve

LAI map

PNC map

PROJECT SATURNO: SUPPORTING VRT FERTILIZATION







28/08/2018 72

SATELLITE: MODERATE RESOLUTION MONITORING

SATELLITE: CROP PHENOLOGY

SATELLITE: PHENOLOGICAL ANALYSIS FROM TIME SERIES

Today

SPOT5-VGT

TERRA-MODISAQUA-MODIS

PROBA-V

1998 2014

SPOT4- VGT1999 2002 2016

300 m 500 m

1000 m

VIS-NIR

SWIR

TIR

250 m 333 m1000 m

500 m 333 m1000 m

1000 m

SpectraIIndices Vegetation growth: NDVI, EVI, etc.- Moisture/water condition: NDWI, NDFI, etc.

EXISTING DATA AVAILABLE: OPERATIONAL SATELLITE SYSTEMS

SENTINEL 3

Daily

Composite SPOT-VGT MODIS PROBA-V Sentinel-3

TIME SERIES ANALYSIS

Treat the signal of a vegetation image image stack to estimate information on plant dynamics

Parameter: (a) beginning of season, (b) end of season, (c) length of season, (d) base value, (e) time of middle of season, (f) maximum value, (g) amplitude, (h) small integrated value, (h+i) large integrated value.

Processing steps: a) Dowload dati, b) pre-processamento, c) calcolo NDVI, d) smoothing del segnale, e) interpolazione giornaliera

a b cd

e

Informazioni sintentiche sulla variabilità spaziale e inter-annuale delle pratiche agricole

MODIS (S2/S3 in future)

CROP MONITORING: SOWING PERIOD AND PHENOLOGY

Informazioni sintentiche sulla variabilità spaziale e inter-annuale delle pratiche agricole

CROP MONITORING: SOWING PERIOD AND PHENOLOGY

Problems for mixed pixel

Very good results when crops entirely cover pixe

Criticalities for field level monitoring

Near real time (NRT) crop monitoring at parcel scale

DOI

LAI

Sowing method DVS

Date

Field measurements

EO retrival

Dry

Water

LAI

Uncertainty

RICE PHENOLOGICAL STAGE OCCURENCE ESTIMATES AT PARCEL LEVEL

Mirco Boschetti; Lorenzo Busetto, Luigi Ranghetti; Francisco Javier García-Haro; Manuel Campos-Taberner; Roberto Confalonieri; TESTING MULTI-SENSORS TIME SERIES OF LAI ESTIMATES TO MONITOR RICE PHENOLOGY: PRELIMINARY RESULTS – IGARSS 2018

A Copernicus services Concept

84

Crop identification

Crop phenology LAI Biotic Risk

Rainfall Maximum humidity

Max Temperature Radiation

COPERNICUS LAI (GEOV1)

WARM MODEL

RRS CROP MONITORING PLATFORM (HTTP://ERMES.DLSI.UJI.ES/PROTOTYPE/GEOPORTAL/LOGIN.HTML)

RRS FROM DATA TO BULLETTINS

LRS: CROP MANAGEMENT PLATFORM (HTTP://ERMES.DLSI.UJI.ES/PROTOTYPE/GEOPORTAL/LOGIN.HTML)

86

Constant Pattern map

Early season fieldhomogenetySAR data (CSK/S1)

Seasonal Pattern mapsOptical data (RE)

Phenological stages Biotic Risk

Documents

Earth Observation systems for operational monitoring of crop conditions · 2020. 5. 25. · 1957 - Sputnik spacecraft ... 400 460 520 580 640 700 760 820 880 940 1000 1060 1120 1180