Cassie Presentation

8/14/2019 Cassie Presentation

1/24

Conversion &

Calibration:

Cassie Knierim

Student Airborne Research


2/24

Overview

vBac kg round In for mati onvImportancevObjec ti ve s

vData C ol lecti onvMAST ER and i n the fi eld

vAnalysisvCon ver ting fr om radi ance to temp er at ur esvCali br at ionvMo de ls to ap pr oxi mate cano pytemp er at ur e

vConclusion


3/24

Importance

vStu die s show th e le af tempe ratu res ofhealthy plants will be a few C lessth an a ir t emper ature .vIn w ater -st re ssed agr icult ur alregi ons , li ke Ke rn Coun ty, CA,eff ici ent and pr eci se irriga tion pl ansare re qu ired.vNumerous indi ces ex ist forde ter minin g rela tiv e p la nt hea lthvmany requi re accur at e cano py


4/24

Local ThermalMeasurements

Advantages:

6. Better accuracy (less atmosphericeffects, etc.)

7. Better spatial resolution

Disadvantages:

12.Small field of view Inefficient for irrigation

planning in larger fields14.Thermal guns are expensive

Remote ThermalMeasurements

Advantages:

5. Increased field of view Better for irrigation plans of

large field7. Free satellite images available

daily

Disadvantages

11.Poor spatial resolution12.More calibration required for

accurate results


5/24

Objective

vCon vert a nd ca li br ate th e rem otelysen sed da ta f rom th e M ASTER i ntoaccu ra te can op y tempe ratu res forth e a lmon d f ield for u se i n stu dy ingcrop h ealth .


6/24

Data CollectionIn the Field

St udy si te: Almo nd Fi eld snear Lost H ills, Califor niaFi eld Day: 7/2 2/0 9

C


7/24


Cali br at ion Si tes :CAL1, CA L2 = Bare Soi lCAL3, CA L4 = W ate rMea su red te mper atur e in a gr id p att ern(~ 10m 2/g rid bo x)

T1 T4 T7

T2 T5 T8

T3 T6 T9

C

C

C

C

1. Re co rd time andlo cat ion.3. Re co rd scale fact orus in g b lackbo dycali brat io n t ar get .5. Measur e tempe rat ur e o feach gridbo x (d id no tus e cent er g rid fo rwat er si tes)

boxe s con tain ing


8/24


Measured the temperature of thecanopy using a handheld thermal gunand a bucket truck.

Measurement Pattern (example):

Location: 226W(truck is on theWest side ofTree 226)

Column 1: Full CanopyColumn 2: Mixed CanopyColumn 3: Soil/ShadeColumn 4: Mixed CanopyColumn 5: Full Canopy

1 2 3 4 5

1

1

1

1

2

2

2

2

3

3

3

3

4

4

4

4

5

5

5

5

22


9/24

AnalysisGeneral Model

MASTER

L

L

(1-)

L

L0= radiance emittedby the atmosphere

Lg = radiance that

reaches surface fromsun

LBB = radiance of

blackbody at sametemperature as surface

= emissivity of


10/24

AnalysisGetting radiance from surface

Atmospheric Correction(MODTRAN 4)

Lg from MODTRAN 4

MASTER

L

L

(1-)

L


11/24

AnalysisGetting temperatures

T = temperature in Kelvin,

= wavelength of the band center in meters

h = Plancks constant (6.62606810-34m2kg/s)

c = speed of light (299792458 m/s)

k= Boltzmanns constant (1.38065031023 m2 kg/s2 /K)

LBB= radiance in W/m2/sr/m


12/24

Analysis

Table 1. Raw data collected in the field on 7/22/2009 usinga handheld thermal gun. Time: Universal Coordinated

Time (UTC). TIR: scale factor for calibration of thermal gun


13/24

Analysis

Table 3. Scale factor applied to Ctemperatures and then converted to Kelvin(+273.15). Algae grids removed. Averages andstandard deviation for each calibration site and


14/24

Analysis

Plot of average calibration temperatures for each site at two

different times. Linear trendline plotted to approximate change intemperature over time (assumed linear change).


15/24

AnalysisCalculating Temperatures from MASTER data

Note: Peak em it tedwavel engt h for Eart hfal ls in to Band 42.


16/24

AnalysisCalculating Temperatures from MASTER data


17/24

Analysis

CalibratedTemperatures appearto be reasonablebased on the fieldmeasurements (pixeltemperature is aweighted average ofsoil and canopy

Temperature Image of

From Field Measurements


18/24

AnalysisResults: Calibrated Temperatures

CalibratedTemperatures

of Almond

Difference

AgainstUncalibratedTemperatures


19/24

Analysis

-3

0

10

Ai r Temp. Pi xel Tem p.Ai r Temp. at Time ofFl igh t: 304. 82 K (fr omBel ri dge Sta ti on )


20/24

Limitations

vField data coll ected on 7/2 2,MASTE R da ta on 7 /24vLimi ted ca nop y t emperatur e

mea su remen ts fo r vali dati onpurposesvPossi ble error in atm osp her iccorrecti onvCali br ati on li ne fr om only 4points


21/24

Conclusion

vCalibrationvCa li br ation wi th fi el d measur ementsis requ ired for accurate canopytemper atu resvMany fa ctors aff ec t th e re su lt sinc ludi ng: a tmosph er ic effe cts,pe rcent c anop y c ove r ver sus per centsha ded and pe rcent sunny soil ,em issi vi tie s, geom et ry of th e c rops,sun a ngle , sen sor loc ati on/ angl e, et c.

vResul tsvThr ee of t he fou r fie ld s had pi xeltemper atu res less tha n a ir

h h


22/24

Further Research

vImpr ove m odels fo r at mo sph er e andchar act eri st ics o f sur facevStud ied a mo del de si gned fo rorang e groves that rep ort edsi gni fi cant accur acy , but lack ofvalues fo r inp ut p ar amet ers andpo ss ib le inap pli cab il it y cause d itto fai lvApp ly unmi xi ng result s (app roxi mat epe rcent ag e of co mpo nen ts in e ach pix el )

vAcqu ir e mo re fi eld dat a fo r imp rov edcali br at ion and mo del vali dati on

A k l d t &


23/24

Acknowledgements &References

Thanks to Shawn & Ni ck, and th e re st ofthe ET grou p for a ll the ir he lp andadv ice! Th anks to ev eryon e who ma deSARP possible !MASTER Wo rk sho p Ha ndboo k, 200 9, Cent ro Nac iona l deAl ta Tecnolo gaA Phy sic al Model fo r Inter pre ting t he La nd S urfa ceTemper atu re Obta ine d by Remo te Senso rs overIn complete C anopies, Vice nt e Ca sell es, Jo se A . So bri no ,a nd C esa r C o l l


24/24

Questions?

Documents

Cassie Presentation