remote sensing & photogrammetry...

Remote Sensing & Photogrammetry

L6

Beata HejmanowskaBuilding C4, room 212,

phone: +4812 617 22 72605 061 510

galia@agh.edu.pl

Image processing

1. Visual interpretation of single spectral band• Readout of DN and coordinates: x,y

2. Image enhancement• Histogram calculation• Linear histogram streatching• Image comparison before and after stretching• Different parameters of linear stretching• Histogram saturation and equalization

3. Visualization of multispectral bands• RGB• FCC

Image processing4. Multi bands operation

• Ratio – Vegetation Index (VI)• Normalized ratio – Normalized Vegetation Index (NDVI)• Multi chanel statistics• Principal Componemt Analysis (PCA) (presented live – Lecture 5)• Map algebra• Image fusion

5. Image classification• Density slicing• piece-wise linear stretchning• Mulispectral image classification

• Sampling• Different algorithms• Classification accuracy assessement• Post classification operation

KB 234

TM4

TM2

TM3

N

VI= TM4 / TM3

Vegetation index

Wykłady TiF II -2010/11

Regina Tokarczyk

Ratio Vegetation Index VI=NIR/REDNDVI Normalized Difference Vegetation IndexNDVI=(NIR-RED)/(NIR+RED)

NDVI – Normalized DifferentialVegetation Index

NDVI=(TM4-TM3)/(TM4+TM3)

Normalized Burn Ratio

• Normalized Burn Ratio

NBR = (TM4 - TM7) / (TM4 + TM7)

• Normalized Difference Burn Ratio (NDBR)

NDBR = NBRpre – NBRpost

Wykłady TiF II -2010/11

Regina Tokarczyk

SI radiometry units

IrradianceIrradinace – ratio of incidient radiant flux on elementar unit flux stosunek strumienia padającego w elementarnej jednostce czasu na element powierzchni odbiornika do wielkości tej powierzchni dla wąskiego przedziału spektralnego p.e.m. (emitancja spektralna)

E = dQ/dt/dA=Φ/dA W/m2

Constant on the top of atmosphere : mean value 1366,1 W/(m2) – solar constant

Eλ

= dQ /dt/dA/ kąt bryłowy W/(m2 sr µm)

AlbedoAlbedo (reflectance coefficient) = Energy reflected fr om the Earth/ Energy incident on the Earth = Q ρλρλρλρλ/Qλλλλ=Lλλλλ/Eλλλλ

(in all EM range)

Albedo depends on:-Enegry incident (solar zenith angle) – on horizontal and sloping surfaces-Reflecting object surface (color, charakter, roughness, wetness)

Lambert’s law:

Energy incident on the Earth = Energy incident on the Eart h (vertical) * cos (Vo)

- Vo – local zenith angle

Example albedo: • Soil 5-10% • Conifer forest 15-20% • Grass 20-25% • Snow: fresh 75-95%, old 50-60% • antopological materials ( 5-20% asphalt, 10-35% concrete, 20-35% stone, 10-35% tile

Map algebra - radiance• Radiance (luminance) calculation:

• Lλ – spectral radinace recorded by sensor• Lmin – minimal radiance of detectors in PAN: – 5.00 • Lmax – maximal radiance of detectors in PAN: 244.00• DN – of channel 8, PAN

( )( ) minminmax 255/ LDNLLoffsetDNgainL +⋅−=+⋅=λ

SESUN

dL

θπρ

λ

λ

cos

2

⋅⋅⋅

=

where:ρ– albedoLλ – spectral radiance recorded by sensord – distance between Earth and Sun in astronomical unitsfor given day of the yearESUNλ – mean irradinace = 1368.00 θS – Sun zenith angle

Map algebra - albedo

Albedo

DNPANBand 8Landsat ETM+

Albedo(0.52 −0.90 µm)

Map algebra - albedo

60%

50%

40%

30%

20%

10%

00,4 0,6 0,8 1,0 1,2 1,4 1,6 1,8 2,0 2,2 2,4

gleba

roślinno ść

wody

B G Rśrednia

podczerwień

długość fali

bliskapodczerwień

um

krzywa spektralna - las li ściasty

0

10

20

30

40

50

60

0 2 4 6 8 10 12 14

długo ść fali [mikrometr]

wsp

ółcz

ynni

k od

bici

a [%

]

Krzywa spektralna - woda

0

10

20

30

40

50

60

0 2 4 6 8 10 12 14

długo ść fali [m ikrometr]

wsp

ółcz

ynni

k od

bici

a [%

]

krzywa spektralna - las li ściasty

0

0.2

0.4

0.6

0.8

1

0 5 10 15

długo ść fali [mikrometr]

wsp

ółcz

ynni

k em

isyj

ności

[-]

krzywa spektralna - woda

0.000

0.200

0.400

0.600

0.800

1.000

0 2 4 6 8 10 12 14długo ść fali [mikrometr]

wsp

ółcz

ynni

k em

isyj

ności

[-]

T

Spectral curves

Eρ (λ) =ρ (λ) E (λ) Es (λ) = ε (λ) B(λ,Ts) + (1-ε (λ)) E (λ)

VIR, SWIR, TIR

EM laws TIR

E=εσT4

σ - Stefan-Boltzmann constant,)/(1067,5 28 KmW ⋅⋅ −

EM laws TIR

Aluminium kinetic temerature100C +273= 283 0K

Powierzchnia wypolerowanawsp. emisyjności 0.06

Powierzchnia pomalowanawsp. emisyjności 0.97

Temperatura radiacyjna(0.06)1/4 283 0K = - 133 oC

Temperatura radiacyjna(0.97)1/4 283 0Κ = +8 οΧ

rzeczTT ⋅= 4 ε

Ε= σΤ4=σ[(ε)1/4Τrzecz]4=σεΤ4rzecz

Radiant temperature

Map algebra - temperature

Wykłady TiF II -2010/11

Regina Tokarczyk

Radiant temperature

• where:

• Lλ – spectral radiance of the sensor, • Lmin – minimum spectral radiance , in

thermal band 0.00, • Lmax – maximum spectral radiance , in

thermal band 17.04

• DN – of thermal band

( )( ) minminmax 255/ LDNLLoffsetDNgainL +⋅−=+⋅=λ( )112 −−− ⋅⋅⋅ msrmW µ

Temperature

Map algebra - temperature

DN band 6 Landsat ETM+

Radiant temperature

Map algebra - temperature

Image processing4. Multi bands operation

• Ratio – Vegetation Index (VI)• Normalized ratio – Normalized Vegetation Index (NDVI)• Multi chanel statistics• Principal Componemt Analysis (PCA) (presented live – Lecture 5)• Map algebra• Image fusion

5. Image classification• Density slicing• piece-wise linear stretchning• Mulispectral image classification

• Sampling• Different algorithms• Classification accuracy assessement• Post classification operation

Image fusion

merging, band sharpening …

+

Multispectral imagecompositionlow spatialresolution

PANHigh spatial

resoluion

=

IHS TRANSFORMATION

KB 123 H SI

Image fusion

KB 123 H SI

Image fusion

IHS TRANSFORMATION

Image fusion

123

174

Image fusion