multi-source spectral approach for early …in.bgu.ac.il/en/epif/site...

MULTI-SOURCE SPECTRAL APPROACH FOR

EARLY WATER-STRESS DETECTION IN

ACTUAL FIELD IRRIGATED CROPS

Maria Polinova1, Thomas Jarmer2, Anna Brook1 1Spectroscopy & Remote Sensing Laboratory Center for Spatial Analysis Research

(UHCSISR), Department of Geography and Environmental Studies, University of

Haifa, ISRAEL - [email protected], [email protected]

Department of Geography and

Environmental Studies

2Remote Sensing Group, Institute of Computer Science,

University of Osnabrueck, Germany - [email protected]

mailto:[email protected]





Optimal irrigation planning

2

Inputs

Crop

Soil properties

Weather

Type of irrigation

etc.

Irrigation

schedule Analysis

Grant #41683‎ - Funded by Israeli Water Authority

Field feedback

3

Crop growth rate: height, LAI,

green biomass, etc.

Canopy scale Leaf scale

Photosynthesis rate, light use

efficiency, chlorophyll and

carotenoid level, etc.

Vegetation features by spectrum

4

Elowitz, M. R. (2013). What is Imaging Spectroscopy (Hyperspectral Imaging)?. Retrieved

November, 27, 2013.

Satellite spectral imagery

5

Benefits Disadvantages

Free or Mid-cost Low resolution (spatial, spectral and

temporal)

Wide spectral range Weather conditions

Big area cover No survey managing

Spectral camera (airborne, field)

6


Big area cover Expensive

High spatial and spectral resolution Weather conditions

VIS-NIR-SWIR range

Survey manage

Field spectrometer

7


Low-cost Point-by-point measurement

High spectral resolution Weather conditions

VIS-NIR-SWIR range

Direct sample measures

Autonomous surveys

Laboratory measures

Vegetation Indices (VIs)

8

VI Development

technologies

Type of

sensitivity Formula Application in agriculture Reference

Normalized Difference

Vegetation Index

(NDVI)

Satellite Green

biomass 𝑤880−𝑤680

𝑤880+𝑤680

Water content, physical parameters

(height, LAI, biomass, dry matter),

predicting yield

Rouse et al.,

1973

Green

Atmospherically

Resistant Index (GARI)

Satellite Chlorophyll

𝑤800 + (𝑤550 − 1.7(𝑤450 − 𝑤670))

Crop coefficient, chlorophyll level,

LAI, soil salinity, predicting yield

Gitelson,

Kaufman and

Merzylak, 1996

Modified Triangular

Vegetation Index

(MTVI)

Airborne LAI

2(𝑤800 − 𝑤550) − 2.5(𝑤670 − 𝑤550

Crop coefficient, chlorophyll level,

LAI, yield predicting

Haboudane et

al., 2004

Leaf area index for

cotton

Field

spectroscopy LAI exp(4.78ln

𝑤900−𝑤680

𝑤900+𝑤680+ 1.55)

Crop coefficient, green LAI,

predicting yield Gutierrez, 2012

Pigment-specific

simple ratio

algorithms

Field

spectroscopy Chlorophyll 19.53

𝑤810

𝑤676

1.48+ 6.65

𝑤810

𝑤676

1.61

Nitrogen and water concentration,

chlorophyll level, pests detecting Blackburn, 1998

Study Area

9

RapidEye Satellite Sensor (5m)

10

Chickpea Tomato & Cotton

5 pixels

Landsat 8 OLI8 (30m)

11

Chickpea Tomato & Cotton

1 pixels

OceanOptics USB4000-VIS-NIR (1cm)

12

10 measures

Dynamics of VIs temporal behavior

13

1. Calculate average means

2. Estimate variability of VIs by coefficient of variability

Cv = s /| x̄ |

where Cv is a coefficient of variation, s is the standard deviation, and x̄ is the

average mean

Tomato

14

May/2015 Jun/2015

Jul/2015

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

5/16/2015 6/5/2015 6/25/2015 7/15/2015 8/4/2015

Plant water concentration

Total water amount=506mm

0

10

20

30

40

50

60

3/15/2015 4/15/2015 5/15/2015 6/15/2015 7/15/2015

Weekly irrigation amount

Tomato

15

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

3/17/2015 5/6/2015 6/25/2015 8/14/2015

NDVI

RapidEye OLI8 Field_spec

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

3/17/2015 5/6/2015 6/25/2015 8/14/2015

GARI


0

0.2

0.4

0.6

0.8

1

1.2

3/17/2015 5/6/2015 6/25/2015 8/14/2015

MTVI


0

500

1000

1500

2000

2500

3000

3500

4000

4500

3/17/2015 5/6/2015 6/25/2015 8/14/2015

PSSR


35% 48%

20%

0

0.5

1

1.5

2

2.5

3

3.5

3/17/2015 5/6/2015 6/25/2015 8/14/2015

LAIc


17% 19% 25%

25%

54% 30%

68%

127%

Cotton

16

May/2

015

Jun/2

015

Jul/2

015

A

ug/2

015

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

4/26/2015 5/26/2015 6/25/2015 7/25/2015 8/24/2015 9/23/2015


Stressed (dry leave)

0

5

10

15

20

25

30

35

40

45

50

5/22/2015 6/22/2015 7/22/2015 8/22/2015

Weekly irrigation amount


Cotton

17

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

3/27/2015 5/16/2015 7/5/2015 8/24/2015

NDVI

RapidEye OLI8 Field_spec Field_stress

-0.1

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

3/27/2015 5/16/2015 7/5/2015 8/24/2015

GARI


0

0.5

1

1.5

2

2.5

3

3/27/2015 5/16/2015 7/5/2015 8/24/2015

LAIc


0

500

1000

1500

2000

2500

3/27/2015 5/16/2015 7/5/2015 8/24/2015

PSSR


19%

0

0.2

0.4

0.6

0.8

1

1.2

1.4

3/27/2015 5/16/2015 7/5/2015 8/24/2015

MTVI


16% 24% 24% 21%

17% 21%

25% 16%

20%

35% 18%

44%

42%

43%

101% 75%

15%

124%

102%

Chickpea

18

Jan/2

015

M

ar/2

01

5

Apr/2

015

M

ay/2

015

Jun/2

015

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

1/16/2015 2/15/2015 3/17/2015 4/16/2015 5/16/2015 6/15/2015


0

5

10

15

20

25

30

35

5/18/2015 5/25/2015 6/1/2015

mm

Daily irrigation amount


Chickpea

19

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

1/6/2015 3/7/2015 5/6/2015 7/5/2015

NDVI


-0.2

-0.1

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

1/6/2015 3/7/2015 5/6/2015 7/5/2015

GARI


0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1/6/2015 3/7/2015 5/6/2015 7/5/2015

MTVI


0

0.5

1

1.5

2

2.5

1/6/2015 3/7/2015 5/6/2015 7/5/2015

LAIc


0

200

400

600

800

1000

1200

1400

1600

1/6/2015 3/7/2015 5/6/2015 7/5/2015

PSSR


30%

15% 18% 20%

47% 26%

34%

30% 30%

18%

26% 35% 80%

19%

30% 23%

37%

25%

58% 88%

35%

38% 75% 43% 24%

50%

Assessment of variability level

20

Test datasets

Cotton Tomato Chickpea Stressed

cotton Chickpea on

22.05.15

30 samples

22.05.15

(flowering)

07.06.15

(flowering/fr

uit

formation)

28.07.15

(ripening)

40 samples

22.05.15

(vegetative)

07.06.15

(vegetative)

28.07.15

(flowering)

31.08.15

(fruit

formation)

30 samples

22.01.15

(emergence)

15.03.15

(vegetative)

19.04.15

(flowering)

18 samples

22.05.15

(vegetative,

with stress

signs)

10 samples

22.05.15

(flowering/

fruit

formation)

Variability in health and stressed crops

21

0%

20%

40%

60%

80%

100%

PSSR

0%

20%

40%

60%

80%

100%

LAIc

0%

20%

40%

60%

80%

100%

GARI

0%

20%

40%

60%

80%

100%

NDVI

0%

20%

40%

60%

80%

100%

MTVI

Cotton

Tomato

Chickpea

Stressed

cotton

Chickpea on

22.05.15

Water-stress detection

22

Satellite imagery Field spectroscopy

Estimation growth rate

(LAI, green biomass,

crop coefficient)

Compression with

reference values

Calculation VIs

(photosynthesis rate,

pigmentation) and their

variability

Assessment the

absolute values

23

Irrigation optimization by spectral data

23

Inputs

Crop

Soil properties

Weather

Type of irrigation

etc.

Irrigation

schedule Analysis

Field feedback

by spectral data

Detected stress by field spectroscopy

Water requirement by spectral image

Thank you for your attention!

Maria Polinova

Spectroscopy & Remote Sensing Laboratory

Center for Spatial Analysis Research (UHCSISR)

Department of Geography and Environmental Studies

University of Haifa

ISRAEL

+972 (0)4 824 96 12 (Office)

+972 (0)4 824 96 05 (Fax)

[email protected]

multi-source spectral approach for early …in.bgu.ac.il/en/epif/site...

Documents