Preliminary After-launch GOCI Characterization: Inter-slot radiance discrepancy issue

Preliminary After-launch GOCI Characterization:

Inter-slot radiance discrepancy issue

Young-Je Park*, Hee-Jeong Han, Seongick Cho, Joo-Hyung Ryu, Jae-Hyun Ahn and Yu-Whan Ahn

Korea Ocean Satellite Center, Korea Ocean Research and Development Institute

Presented at IGARSS 2011, Vancouver, Canada

Objectives

• To understand the inter-slot radiance discrepancy issue

• To seek ideas/suggestions on how to approach this GOCI specific issue

Sensors have specific issues

• MODIS stripe noise– detector calibration– difference in mirror side characteristics– Sensitivity to polarization state

• MERIS– SMILE effects: wavelength variation – Discontinuity at some camera interface

Outline

• Overview of the GOCI optical system and image acquisition sequence

• Inter-slot discrepancy: – variability within a slot– variability across different slot boundaries– variability with observation hours (0, 3, 7 hours)– How RT simulations show

• Image smoothing technique• Future directions

GOCI sensor

GOCI optical layout

Three Mirror Anastigmatic Telescope

GOCI slots imaging sequence

Target Area

20,00

25,00

30,00

35,00

40,00

45,00

50,00

110,00 115,00 120,00 125,00 130,00 135,00 140,00 145,00 150,00Longitude (deg)

Lat

itude

(deg

)

P1 (+Y,-Z)P2 (+Y,+Z)

P3 (-Y,+Z)

P4 (-Y,-Z)

Slot 1

Slot 16

Slot 4

Slot 5Slot 8

Slot 9 Slot 12

Slot 13

P6

P5P7

P8

1 2 3 4

56

16

9

8 7

15 14 13

121110

Imaging procedure for a GOCI slot

Nominal time intervals for GOCI operation

• Interval between bands = ~ 8 seconds• Interval between consecutive L1a slots = ~ 103

seconds• Duration for acquiring one GOCI image = ~ 103*16

seconds = 27 minutes• Interval between consecutive GOCI images = one hour• Interval between the adjacent slots in L1B scene = up

to ~103*7 seconds or 12 minutes => sun angle difference??

Requirements for comparing radiances from two slots

• Accurate geometric registration• Spatially homogenous conditions for the

atmosphere and water are preferred, which is to avoid seeing different air/water mass from two different slots

Inter-slot discrepancy

Variability within a slot

• 20110330_0h image: slot 3-6 border

Variability across different slot boundaries

• 20110330-3h

Inter-slot discrepancy: spectral aspect(033003)

• Slot #2-7 border

Inter-slot discrepancy: spectral aspect(033003)

• Slot #3-6 border

• Slot #4-5 border

• Slot #5-12 border

• Slot #6-11 border

• Slot #7-10 border

• Slot #8-9 border

• Slot #9-16 border

• Slot #10-15 border

Variability with observation hour

Slot border reflectance change

• Within a slot border: moderately variable with consistent difference spectra

• For different slot borders: variable magnitude, moderately variable spectra

• For different observation hours: larger difference (lower reflectance for the upper slot) in the 7h image

• Bands 7 & 8 reflectance ratio changes significantly, which has a serious effect on atmospheric correction that uses those bands.

How does RT code simulate the discrepancies?

uslot=3,lslot=6: 3099,1584 3099,1585lat,lon= 41.0020866 131.8832245sunz= 54.8425102 54.0042229suna= 121.0095673 122.2263718senz= 47.5428658 47.5378571sena= 185.6037445 185.6040649

uslot=3,lslot=6: 3099,1592 3099,1593lat,lon= 40.9659882 131.8822021sunz= 37.4894829 37.5132179suna= 180.6660614 182.7969666senz= 47.5027428 47.4977188sena= 185.6062317 185.6065369

uslot=3,lslot=6: 3099,1597 3099,1598lat,lon= 40.9434242 131.8815613sunz= 65.5979691 66.5323486suna= 252.3380127 253.3214874senz= 47.4776611 47.4726372sena= 185.6077881 185.6080933

• 0h: 9hr local time

• 3h: 12hr local time

• 7h: 17hr local time

Simulation with AOT550=0.1

400 450 500 550 600 650 700 750 800 850 900-0.004

-0.003

-0.002

-0.001

0

0.001

0.002

0h3h7h

wavelength (nm)

TOA

refle

ctan

ce d

ifere

nce

Simulation with AOT550=0.5

400 450 500 550 600 650 700 750 800 850 900-0.004

-0.003

-0.002

-0.001

0

0.001

0.002

0h3h7h

wavelength (nm)

TOA

refle

ctan

ce d

ifere

nce

GOCI data

400 450 500 550 600 650 700 750 800 850 900-0.008

-0.006

-0.004

-0.002

0

0.002

0.004

0.006

0h3h7h

wavelength (nm)

TOA

refle

ctan

ce d

ifere

nce

An image smoothing technique

• Distance-to-border weighted average– Applied to overlapped area– Simple and good for image generation– Smoothing the TOA reflectance data will not be

good for downstream data processing including the atmospheric correction.

– Smoothing the geophysical parameters would make sense.

Distance-to-border weighted average

d1

Slot i

d2

d3

d4

Slot i

wi

wi = min(d1,d2,d3,d4) where is number of pixels to the k-th border

Slot j

wj

N’=∑(wi Nⅹ i)/∑wi

N’: weighted averageNi: reading from the ith slot

Example 1 (original)GOCI 20110412-07h, South Japan

Example 1 (weighted average)GOCI 20110412-07h, South Japan

Future work

• Clarify questions of – Is it an issue of the GOCI radiometric calibration?– Is it an issue of the band filter properties?– Is it an issue of the ghost image?

• Develop a scientifically based model to correct the inter-slot discrepancy. Bands6,7,8 are critical for atmospheric correction.

Thank you!

Please contact us if you have any idea on this issue.

youngjepark@kordi.re.kr