jordi bort- spectroradiometric approaches to investigate dro

8/2/2019 Jordi Bort- Spectroradiometric Approaches to Investigate Dro

1/89


2/89

Average maize yields in the US

1860 1880 1900 1920 1940 1960 1980

Year

0

2

4

6

8

10

Grain Yield (t/ha)

Open PollinatedPopulations

Doubled

Hybrids

F1 Hybrids

Average maize yields in the US

1860 1880 1900 1920 1940 1960 1980

Year

0

2

4

6

8

10

Grain Yield (t/ha)

Open PollinatedPopulations

Doubled

Hybrids

F1 Hybrids Maize US 1860-1990

Soybean US 1940-2000

Wheat 1866-1996

Genetic gains of major crops

Chrispeels MJ & DE Sadava.

Plants, Genes and AgriculturePlants, Genes and Crop BiotechnologyJones and Bartlett Publ.


3/89

Yield production and drought stressYield production and drought stress

Redrawn from Calderini & Slafer 1998

Argentina

Australia

Canada

France

Germany

UKUSA

Countries where wheat is

mostly grown undermoderate to little or no

severe drought drought

0

2

4

6

8

0 20 40 60 80

Years (from 1900)

Grainyie

ld(Mgha

-1)

0

24

6

0060 80

100

Barley

Argentina

Australia

Canada

France

Germany

UKUSA

Countries where wheat is

mostly grown undermoderate to little or no

severe drough

t drought

0

2

4

6

8

0 20 40 60 80

Years (from 1900)

Grainyie

ld(Mgha

-1)

0

24

6

0060 80

100

Barley


4/89

How increases in yield potential have

been attained in the past?

Breeding gains have been achieved selecting by yield itself

as the main (and frequently) the only trait evaluated

Shorter plants with a higher HI have been the responsibleof the increase in yield potential

Under drought this approach is complicated by theexistence of important GxE interactions and the higherwithin-site variability that also diminishes heritability

Physiological traits have been seldom used in the past astrue selection criteria The main reason was the difficulty oftheir measure on practical breeding programs


5/89

drought is responsible for most losses in rainfed agriculture,particularly on Mediterranean climates

drought is frequently a combination of water, temperature

and radiation stressesdrought is the most widely stress studied (and then withmore information available to reach sound general

conclusions)

Physiology-aided breeding for stress

environments


6/89

Genetic improvement in this context may be approached

through selection either:

EMPIRICAL : directly, for a primary trait (normallyyield) under the targeted environment (with the

specific stresses naturally occurring; Ceccarelli &Grando, 1996), or

ANALYTICAL or PHYSIOLOGICAL: indirectly, for asecondary trait, that must be putatively related to animproved behaviour of the crop when it is grown in astressful environment

Physiology-aided breeding for stress

environments


7/89

The putative secondary traits for an analytical breedingprogram may be used:

Identifying prospective parents for crossing, independentlyof whether the subsequent selection is to be made byprimary (e.g. yield) or secondary traits

As a direct selection criteria in segregating generations

Physiology-aided breeding: secondary traits


8/89

Regardless the actual contribution of analytical breeding, mostbreeders develop a profound understanding of their environments

and adaptation of their genetic materials.Physiological assessment of adaptation to the environment isneeded to complement breeders' impressions particularly in thefirst and late stages of a breeding program.

A germplasm strategy is also needed for breeding for drought aswell as for any other trait. Most breeders focus on just the elitegene pool, reflecting decades of crossing, selection and

recombination. In fact there is a significant gap between the eliteant the unimproved gene pools.

As empirical breeding seems to be reaching a plateau we mayneed different approaches to further improve grain yields.

Analytical and Empirical Breeding


9/89

As crops experience in most growing areas the effect ofabiotic stresses, at least during part of its growing season,these approaches must improve the response of the crop tothese stresses.

In this context, more emphasis should be given to the useof new genetic variability particularly through the genetically

building of new parent for crosses, incorporating desiredtraits into the gene pool after a series of pre-breedingactivity.

The value of local landraces of many crops in breeding

programs for dry lands should not be underestimated

Development of modern apparatus and new analytical toolswill facilitate measurement ofphysiological traits in the field.

Analytical and Empirical Breeding


10/89

Yield

Physiology

Growth

Development

Carbon economy

Nitrogen economyWater economy

Physiological Plant Breeding

Photoperiod

Vernalization

Intrinsic earliness


11/89

DroughtDrought -- adaptive traits in C3 cerealsadaptive traits in C3 cereals

1. Early Growth. Pre-Grain filling Early vigor/ground cover Stem carbohydrates reserve

2. Access to water High relative leaf water content Low canopy temperature Osmotic adjustment

3. Water use efficiency

High Harvest index Spike/awn photosynthesis Low 13C

4. Photo-protection Leaf morphology

Pale color Wax/pubescence Posture/rolling

Mathew Reynolds, CIMMYT


12/89

Physiological tools Integrative

- Carbon Isotope Discrimination (13C)

- Canopy Temperature Difference

- Spectroradiometrical Reflectance

- Fluorescence

- Near Infrared Reflectance Spectroscopy (NIRS)

- Ash content

Others

- Phenology- Chlorophyll fluorescence- SPAD and SLDW


13/89

Selection by Secondary Traits

How to choose

a trait ?

How to

evaluate it ?

Requirements

and implications ?Usefullness in Breeding

and Crop Management

What traits

should be used ?


14/89

Selection by Secondary Traits

Limitations

of molecular

or biochemical

aproaches

Yield is a

quantitative

character

Basic

Determinants

of Yield ?

Integrative trait

Genetic correlation

with yield

Heritability > yield

Directly related

with yield

Productivity

or survival ?

Stress scape,

avoidance or

tolerance ?

Negative

interaction

GxE

Ecological approach

Previous definiton

of target environment

How to choose

a trait ?


15/89

Selection by Secondary TraitsSelection by Secondary Traits

Limited utility

of traditionalphysiological

methods

Emergency

of alternativemethods

(ecophysiological...)

Spectroradiometrics

Canopy Temperature

Fluorescence

Remote Sensing

13C/12C and

its surrogates

18O/16O

Stable Isotopes ............ .............

Quick

Easy

Non-destructive

Low cost

How to evaluate a trait ?


16/89

Plant Breeding

Fertilization

Seed density

Phenology Adjustment

Chemical control

Mecanization

Crop Rotations

Manure

Organic matter

Pests

Soil Erosion

Other negatives

0 10 20 30 40 50 60-10-20-30

Relative Contribution (%)

Agriculture and the Environmental Challenge


17/89

Response to stress

Accumulation of ABA (enhance survival but reduces

productivity => tolerance versus avoidanceProtection cellular membranes

Ability of the plant to capture water from a drying soilthrough

a) deeper and/or more thorough root explorationb) through osmotic adjustment

Delay senescence

Many of these mechanisms favor survival but mayhave limited value in enhancing grain yield under avarying levels of water stress

Bruce et al. 2002 J. Exp. Bot. 53: 13-15


18/89

Understanding of crop responses to

environment =>Ideotype approach to plant breeding

Ideotype: plan of the phenotype of a cultivarthat will perform optimally in a specific set of

climatic, soil, biotic and socio-culturalconditions (Hall 2001)

(Hall 2001)


19/89

Ideotype

Yield potential is important in determining yieldunder moderate stress, with yields beyond50% of potential

Grain yield is normally highly correlated withkernel number per unit area and per plantrather than with weight per kernel =>

Factors affecting grain set under drought



20/89

Ideotype

Under Water Stress

High grain yield

Small ASI

Stay green

Under well-watered conditions

Adequate yield

Small tassels Upright leave



21/89

Changes associated with selection

Reduced barrenness under drought (associated with

more rapid ear growth Not too increased biomass production

Slightly earlier anthesis date

Root biomass in the top 50 cm of soil declined by35%, but there was not change in any trait indicativeindicative of plant water status.

Reduced number of spikelets per ear (=> more earlyvigorous silking under drought) => more successful in

forming grain under drought at flowering These mechanism leading to improved yield under

drought also appear to hold under low N



22/89

Remote sensing techniquesRemote sensing techniques

Canopy temperature

Spectroradiometrical Reflectance Indices


23/89

Canopy TemperatureCanopy Temperature


24/89

Canopy Temperature DepressionCanopy Temperature Depression

CTD = TCTD = T airair-- TT plantplant

16.4 a16.1 b15.1 c13

C

2.16a1.94 ab1.68bCTD (C)

ModernVarieties

Old VarietiesLandraces

Landrace Old Varieties Modern


25/89

CTD and YieldCTD and Yield

____Correlation of CTD with yieldAerial Hand-held

Trial n Phenotypic Genetic Phenotypic Genetic

RILs (Seri82*7C66)81 0.40** 0.63** 0.50** 0.78**

Advanced lines 58 0.34** - 0.44** -

**statistical significance at 0.01 level of probability- genetic correlations not calculated due to design restrictions

Reynolds etal., 1999


26/89

Results of a stepwise regression with grain yield as dependent variable and the combination ofphysiological traits as independent variables performed across 12 trials assayed. Traits studied were:DISCR kernel carbon isotope discrimination; TKW, thousand kernel weight; CTDA, canopy temperaturedepression at anthesis; CTDM, CTD at milk grain stage; C, kernel carbon content).

0

10

20

30

40

50

60

70

80

90

DISCR TKW CTDA CTDM C

Percentage

Percentage of environments where thetrait significantly entered the model

Percentage (mean across environments)of model-explained yield attributed to thetrait

Royo et al. 2002 (Aust. J. Agric. Res)

Integrative BreedingIntegrative Breeding


27/89

SpectroradiometricalSpectroradiometrical ReflectanceReflectance IndicesIndices

Different levels of assessment:

- Canopy

- Seedlings- Leaves


28/89


Remote determination by instantaneous, non-invasivemethods of the pigment content of leaf canopies, the

status of these pigments and the crops ability tointercept radiation and photosynthesize is a good way ofdetecting physiological status and stresses in plants.

The use of portable narrow-bandwidth visible/nearinfrared spectroradiometers provides a lot of informationthat can be summarized in a set of indices calculatedfrom formulations based on simple operations betweenthe reflectances at specific wavelengths, such as ratiosand differences.


29/89

H

A

Visible (VIS) Near Infrared (NIR)

Blue Green Red

M

PM

Soil


0

0.1

0.2

0.3

0.4

0.5

300 500 700 900 1100Wavelength (nm)

Reflectance


30/89

Wavelength, nm400 500 600 700 800 900 1000

Refle

ctance

0.0

0.2

0.4

0.6

0.8

1.0

Control

N-deficient

Spectroradiometrics and Nitrogen Status


31/89

SpectroradiometrySpectroradiometry and Green Biomassand Green Biomass

Wavelength, nm400 500 600 700 800 900

Reflecta

nce

0.0

0.1

0.2

0.3

0.4

0.5

0.6

Irrigated

Rainfed-2

Rainfed-1

Bare soil


32/89

Spectroradiometrical Indices

Some indices for remote sensing of crop status.Physiological

parameter

Radiometric

Index

NDVIR R

R R

NIR d

NIR red

=

+

Re

Leaf area, [Chl],

Green Biomass, etc. SRR

RNIR red=

Water Content WIR

R=

900

970

Are they able to detect true genotypic differences,or are they only valuable to discriminate across

major environmental effects?


33/89

Spectroradiometrical Indices

Some indices for remote sensing of crop status.

Physiological parameter Radiometric Index

NDVI

R R

R R

NIR d

NIR red=

+

Re

Leaf area,

[Chl],Green Biomass, etc.

SRR

RNIR

red

=

SAVIR R

R R L

LNIR d

NIR red

=

+ +

+Re

( )1

(where L=0.5 for most crops)

Chl degradationNPQI

R R

R R=

+

415 435

415 435

Car/ChlSIPI

R R

R R=

+

800 435

415 435

PRUEPRI

R R

R R=

+

531 570

531 570

Water ContentWI

R

R=

900

970


34/89

Yield Components

General determinants Yield = IR x AR x PE x HI

IR, Incident Radiation

AR, Absorbed Radiation

PE, Photosynthetic Efficiency HI, Harvest Index

In water-limiting conditions (Passioura 1977) Yield = W x WUE x HI

W, Water used

WUE, Water Use Efficiency

HI, Harvest Index

S di i lS t di t i l i di ii di i t


35/89

0.2

0.4

0.6

0.8

1.0

0 2 4 6 8 10 12

LAI anthesis

NDVI

y0.5

= 0.95 - 0.68 e-x

r 2 = 0.93**

0

2

4

6

8

10

0.2 0.4 0.6 0.8 1.0NDVI

Yield(t/ha

y = 92 + 116 exp (x / 0.23)

r 2 = 0.86**

0

2

4

6

8

10

0.7 0.8 0.9 1.0 1.1

WI

Y

ield(tm/ha

y0.5

= 27 - 363 x2

lnxr 2 = 0.93 ***

SpectroradiometricalSpectroradiometrical indices across environmentsindices across environments

0

1

2

34

5

6

7

8

9

0 5 10 15 20 25 30 35

SR

Yield(t/ha)

y = -3579 + 3258 x 0.34

r2

= 0.86**

Royo et al. 2003. Int. J. Remote Sensing, 24:1-16

Aparicio et al. 2002. Crop Sci., 42: 1547-1555


36/89

WIWI vs.vs. waterwater statusstatus

Relationship between WI andeither grain yield, carbon isotopediscrimination (13C ) and canopytemperature difference betweenthe canopy and the air (T) forbarley cultivated under differentlevels of salinity.

Peuelas et al. 1996


37/89

Measuring spectral reflectance of crop

canopies


38/89

Handheld instrument

NTech Industries, Inc.

740 South State StreetUkiah, CA 95482

GreenSeeker Hand-Held Unit

f f


39/89

Measuring spectral reflectance of

crop canopies


40/89

Spectroradiometrics for individual plants

Spectrorradiometer

Tube with reflecting walls

Halogen lamp

Fiberoptic

Diffuser

Remote Cosine Receptor

Aluminium foil


41/89

Potted DW plantsBIOMASS / NDVI

y = 79,802x - 8,6683

R2

= 0,8845

0,00

10,00

20,00

30,00

40,00

50,00

60,00

0,000 0,100 0,200 0,300 0,400 0,500 0,600 0,700 0,800 0,900

NDVI

SHOOT(gr.DW)

E l i / dE l i / d


42/89

Early vigor/ground coverEarly vigor/ground cover


43/89

Stress Management


44/89


45/89


46/89


47/89

1. Estimate N response in-season.1. Estimate N response in-season.

45 N Preplant45 N Preplant45 N Preplant

90 N Preplant90 N Preplant90 N Preplant

RINDVI = 1.46RIRINDVINDVI = 1.46= 1.46

SolutionsSolutionsSolutions

N stress Management


48/89

Management Solutions1. Estimate N response in-season.

29 Locations, 1998-2002

y = -0,22x2

+ 1,86x - 0,50

R2

= 0,68

0

1

2

3

4

0 1 2 3 4 5

RINDVI

RIHarves

t


49/89

0

1

2

3

4

5

6

0 0.002 0.004 0.006 0.008 0.01

INSEY=NDVI/Days f rom planting to sensing GDD>0

Grainyield,

Mg/h

a

Perkins N&P, 1998

Perkins S&N, 1998

Tipton S&N, 1998

Perkins N&P, 1999

Experiment 222, 1999


Efaw AA , 1999

Experiment 801, 1999Experiment 502, 1999

Perkins N&P, 2000



Efaw AA , 2000



Hennessey, AA, 2000

Perkins N&P, 2001



Efaw AA , 2001



Hennessey, AA, 2001

y=0.4593e246.3x

R2=0.55

YP0YP0

Management Solutions2. Provide in-season estimate of yield(INSEY)

YPNYPN

YPMAXYPMAX


50/89

3. Measure and treat spatial variability, in-seasonManagement Solutions


51/89

Management: Conclusions

Temporal variability can be managed. Create N-Rich Strip in each field.

Evaluate yield potential and N responsiveness in-season using sensor.

Spatial variability can be managed on a fine

resolution (


52/89

Use of Spectral Reflectance inBreeding

Some case studies

Use of Spectral Reflectance in


53/89


Breeding Early prediction of crop yield can be an

important tool for identifying promisinggenotypes in breeding programmes.

Some results indicate that, for durum wheat,

milk-grain stage is the most appropriatedevelopment stage for yield assessment.

However, some indices are also sensitive to

yield variations when determined at anthesis oreven heading or booting.

C. ROYO, N. APARICIO, D. VILLEGAS, J. CASADESUS, P. MONNEVEUX and J. L. ARAUSUsefulness of spectral reflectance indices as durum wheat yield predictors under contrasting Mediterranean conditions

INT. J. REMOTE SENSING, 20 NOVEMBER, 2003,VOL. 24, NO. 22, 44034419



54/89


Breeding





55/89


Breeding The capacity of spectral reflectance indices to

forecast grain yield in durum wheat increased onlocations that allowed genotypes to express theiryield potentiality.

Assessment of differences between genotypesin specific environments reduced the percentageof yield variability explained by those indices.



Association between canopy reflectance indices and


56/89

yield and physiological traits in bread wheat under

drought and well-irrigated conditions

The relationship of SR indices with grain yield and biomass fittedbest with a linear model. NDVI and GNDVI showed positiverelationships with grain yield and biomass under well-irrigated

conditions (r= 0.350.92), whereas NDVI showed a strongerassociation with yield under drought conditions (r= 0.54).

The 6 chlorophyll indices showed significant association with yieldand biomass of wheat genotypes grown under well-irrigatedconditions (r= 0.390.90).

WI showed a significant relationship with grain yield in wheatgenotypes grown under drought stress conditions (r= 0.60) as wellas with grain yield and biomass under well-irrigated conditions (r=0.520.91).

The relationship between WI and CTD was significant (P 0.05) in

both environments (r= 0.440.84). In conclusion, the SR showed potential for identifying higher-yielding

genotypes in a breeding program under dry or irrigated conditions,as well as for estimating some physiological parameters.

Mario Gutirrez-Rodrguez, Matthew Paul Reynolds,, Jos Alberto Escalante-Estrada and Mara Teresa Rodrguez-GonzlezAssociation between canopy reflectance indices and yield and physiological traits in bread wheat under drought and well-irrigated conditions

Australian Journal of Agricultural Research 55(11) 11391147 (2004)

Yield predicting attributes of


57/89

Yield predicting attributes of

spectral reflectance indices Photosynthetic area indices and senescence

indices were good indicators, of biomass and

phenology, respectively, when comparing awetter site with a drier site.

When crop development rate was acceleratedby growing plants under high temperature,provided by a spring-sown trial underMediterranean conditions, all spectral indicesshowed significant variation within a period of

one week through grain filling, reflecting thechanges in crop phenology and the onset ofsenescence.

J. BORT, J. CASADESUS, M. M. NACHIT and J. L. ARAUS

Factors affecting the grain yield predicting attributes of spectral reflectance indices in durum wheat: growing conditions, genotypevariability and date of measurement

International Journal of Remote Sensing Vol. 26, No. 11, 10 June 2005, 23372358


58/89


59/89

Yield predicting attributes




Yield predicting attrib tes


60/89

Yield predicting attributes

Both the sign of the correlation coefficients betweengrain yield and some spectral reflectance indices, and

the changes of those signs throughout the grain fillingperiod of durum wheat canopies, were tracking thecapacities of those canopies to obtain higher yieldsthrough adjustment of phenology, mainly by avoidance

of late grain filling temperatures and drought. Spectral reflectance data give clues to understand which

phenological characteristics of durum wheat canopiescan be selected to improve yield. The results also

illustrated how important it is to define carefully the dateduring the crop cycle when spectral reflectance is to bemeasured.




Spectral Reflectance Indices


61/89


SRI have been suggested as indirect selection criteria byreporting genetic variation for SRI among genotypes, the

effect of phenology and year on SRI and their interactionwith genotypes, and the correlations between SRI andgrain yield and yield components of wheat.

A clear trend for higher association between grain yield

and the NIR-based indices was observed at heading andgrain filling than at booting. Overall, NIR-based indiceswere more consistent and differentiated grain yield moreeffectively compared to the other indices. The results

demonstrated the potential of using SRI as a tool inbreeding programs for selecting for increased geneticgains for yield.

M. A. Babar, M. P. Reynolds, M. van Ginkel, A. R. Klatta, W. R. Raun and M. L. StoneSpectral Reflectance Indices as a Potential Indirect Selection Criteria for Wheat Yield under IrrigationCrop Sci 46:578-588 (2006)



62/89


Spectral indices as a selection tool in plant

breeding could improve genetic gains for

different important traits:

estimate genetic variation for in-season biomass

production, leaf chlorophyll, and canopy temperature

(CT) in wheat (Triticum aestivumL.) under irrigatedconditions.

the potential of using SRI as a breeding tool to select

for increased genetic gains in biomass andchlorophyll content, plus for cooler canopies.

M. A. Babar, M. P. Reynolds,*, M. van Ginkel, A. R. Klatt, W. R. Raun and M. L. StoneSpectral Reflectance to Estimate Genetic Variation for In-Season Biomass, Leaf Chlorophyll, and Canopy Temperature in WheatCrop Sci 46:1046-1057 (2006)

GY vs NDVI phenology(1)


63/89

GY vs NDVI, phenology(1)

IDUWUE Gimenells RILs 2005

R2

= 0,2238

R2

= 0,3254

R2 = 0,041

0

500

10001500

2000

2500

3000

0 0,2 0,4 0,6 0,8 1

NDVI

G

Y22/04/2005

10/05/200531/05/2005



64/89


IDUWUE Gimenells RILs 2005

R2 = 0,1969

R2

= 0,0979

R2

= 0,2413

0

500

10001500

2000

2500

3000

-0,2 0,0 0,2 0,4 0,6

Decrease in NDVI

G

YApril-June

April-MayMay-June



65/89


Spectral reflectance data may help to

understand phenological characteristics ofdurum wheat canopies, such as crop

duration, provided the date of

measurement is well chosen.


66/89

Digital photography as a

screening tool for cerealbreeding.

Jaume Casadess1, Jordi Bort2 and Jos Lus Araus2.

1Institut de Recerca i Tecnologia Agroalimentries (IRTA), Spain.

2Dept. Biologia Vegetal, Universitat de Barcelona, Spain.


67/89

IntroductionDigital Cameras are very popular devices that can be useful

for field data acquisition.

-affordable-portable

-ease of use

Analysis of digital images can bring in many variables,

allowing-objective estimation of some vegetation traits

-gathering of data for statistical analysis

Numerical representation of color


68/89

IHS

Intensity, Hue, SaturationPractical for image analysis

0

120

240

Hue wheel:

Numerical representation of color

RGB: related with color reproduction by computer screens, etc.

CIE-XYZ~ sensitivity of human visual system

Consistent distance

practical for arithmetics

CIE-Lab

There are a number of different systems for representing a given color.All them use 3 quantities (different meaning for each system)

Color data from each image


69/89

Average color:RGB 92.0 91.3 45.8

g

RGB 92.0 91.3 45.8

XYZ 0.096 0.109 0.033

IHS 0.299 59.3 0.401

CIE-Lab 37.1 -10.0 31.233.2-18.140.232.1-11.038.731.2-10.037.1CIE-Lab

0.36375.20.3300.39460.70.3140.40159.30.299IHS

0.0450.1410.1140.0370.1210.1050.0330.1090.096XYZ

53.6106.692.248.596.195.545.891.392.0RGB

Green AreaSoil coverField of view

Beyond averages: histograms for color components


70/89

Histogram of Hue

0

0.05

0.1

0.15

0.2

0.25

2 32 62 92 122 152 182

Hue (0-360)

frequency

Histogram of Hue

0

0.02

0.04

0.06

0.08

0.1

0.12

2 32 62 92 122 152 182

Hue (0-360)

freq

uency


71/89

Howeverunsteady yield of conventional digital cameras.

Due to the cameras self-adjustments, the same object can be recorded

with different colors depending on the general brightness of the scene.

Alternative ways to cope with


72/89

cameras self-adjustments.

Fix the settings and characterize the camera in laboratory

Use reference panels to recalculate the colors.

Select robust parameters, least affected by self-adjustments. R, G, B, I, H, S, X, Y, Z, L, a, b,...

Robustness of different color parameters

A color chart (24 known colors) was recorded over 15 turfs of different color


73/89

Red

0

60

120

180

240

300

0 60 120 180 240 300actual value

recordedvalue

Green

0

60

120

180

240

300

0 60 120 180 240 300actual value

recorde

dvalue

Blue

0

60

120

180

240

300

0 60 120 180 240 300actual value

recordedvalue

Intensity

0

0.2

0.4

0.6

0.8

1

0 0.2 0.4 0.6 0.8 1actual value

recordedvalue

Hue

0

60

120

180

240

300

360

0 60 120 180 240 300 360actual value

record

edvalue

Saturation

0

0.2

0.4

0.6

0.8

1

0 0.2 0.4 0.6 0.8 1actual value

record

edvalue

A color chart (24 known colors) was recorded over 15 turfs of different color.


74/89

Digital photography in breeding programs

AffordableEasy to use

Ubiquitous

Can allow unexpensive field data acquisition

simultaneously at different sites.

Objectives


75/89

Objectives

1. Derive vegetation indices from color

analysis of digital images.

2. Explore the potentialities beyond

vegetation indices: image features that

can contribute to the assessment of

physiological traits.

Materials and methods


76/89

Pictures with conventional

digital cameraNDVI measurement with

GreenSeeker

As many field trials aspossible, showing a wide

range of environments

One NDVI mesurementvs.

One digital picture

Performance of Hue as a Vegetation Index


77/89

e o ce o ue s Vege o deNDVI vs Hue

R2

= 0.90

30

40

50

60

70

80

90

100

110

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

NDVI

Hue

(0-360)

Irrigated

Rainfed

DrySev.Dry

Performance of %Green Area as a Vegetation Index


78/89

NDVI vs. Green Area

R2

= 0.91

0

20

40

60

80

100

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

NDVI

GA,%

Irrigated

Rainfed

DrySev.Dry

Performance of a* as a Vegetation Index


79/89

NDVI vs a* (from CIE-Lab color space)

R2 = 0.87

-35

-30

-25

-20

-15

-10

-5

0

5

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

NDVI

a*

Irrigated

Rainfed

DrySev.Dry

Results


80/89

Example of NDVI and Hue covariation along a range of barley rainfed plots

30

40

50

60

70

80

90

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51

plot number

Hue,

0-360

0.2

0.3

0.4

0.5

0.6

0.7

0.8

NDVI

Hue NDVI

NDVI vs color parameters in more trials


81/89

iduwue RILs Gimenells 050510

30

34

38

42

46

50

0 50 100 150 200 250 300num plot

HU

E

0,1

0,15

0,2

0,25

0,3

0,35

NDVI



82/89

iduwue RILs Gimenells 050531

31,0

32,0

33,0

34,0

35,0

36,0

37,0

0 50 100 150 200 250 300num plot

HU

E

0

0,05

0,1

0,15

0,2

0,25

ND

VI



83/89

iduwue Gimenells RILs 050510

R2 = 0,3254

R2 = 0,2354

0

10

20

30

40

50

60

0 500 1000 1500 2000 2500 3000

GY

HUE

0

0,1

0,2

0,3

0,4

0,5

0,6

ND

VI



84/89

iduwue Gimenells RILs 050531

R2

= 0,041

R2

= 0,0197

30

31

32

3334

35

36

3738

0 500 1000 1500 2000 2500 3000

GY

HUE

0

0,1

0,2

0,30,4

0,5

0,6

0,70,8

ND

VI


85/89

NDVI vs color parameters on each trial.

0.870.910.90all

0.600.670.62Barley arid (later)

0.620.710.67Barley arid

0.770.850.82Durum wheat-Rainfed

0.020.050.06Durum wheat-Irrig.

a*%GreenAreaHuetrial

(R2 for the relationship between NDVI and each color parameter)

Beyond vegetation indices


86/89

Other parameters could be estimated from digital images.

Total soil cover(green+dry vegetation)

Physiological status

(N-content, Chl,...)?

from the color of the

green area only.

Conclusions


87/89

Some color parameters derived from image analysis that

can be used as Vegetation Indices are: Hue, %Green

Area and a* (from CIE-Lab).

At least in the essayed sites, dry vegetation can be

distinguished from the soil and quantified.

These color parameters can be calculated separately for3 regions of the image : total field of view, soil cover and

green vegetation.

The image analysis was performed automatically at arate of 2.2 images / s, with a plain desktop PC.

As many ants as you may eventually find in this field are the many

hours, you need

NOTto be there measuring ecophysiology


88/89

Many Thanks!

NOT


89/89

jordi bort- spectroradiometric approaches to investigate dro

Documents