Development of a Radiometric Correction Algorithm for OCTS Data

Development of a Radiometric Correction Algorithm for OCTS Data

Mitsuharu Tokunaga
Institute of Industrial Science
University of Tokyo
E-mail : toku@shunji.iis.u-tokyo.ac.jp

Keywords : ADEOS, OCTS, RADIOMETRIC CORRECTION

Abstract
A software package for ADEOS/OCTS and NOAD/HRPT Data Analysis (PaNDA/PANDA) has been developing by PaNDA committee consisted from Japanese scientists since 1989. The PaNDA and PANDA are a software package for NOAA/HRPT and ADEOS/OCTS respectively. The PaNDA and PANDA has been distributed to scientists as a freeware. Recently, we have additionally developed the program of the radiometric correction for ADEOS/OCTS data. In this paper, we introduce the radiometric correction algorithm adopted in PANDA software.

1. Introduction
Many commercial software packages for remote sensing are now available, but most of them are expensive for ordinary scientists. Also, few packages reveal the detailed processing algorithm and the source code. By the reason, some Japanese scientists started to develop the software for processing NOAA/HRPT data as a freeware. The package named a software the package named a software package of ADEOS/OCTS and NOAA/HRPT Data Analysis (PaNDA/PANDA). Recently, PaNDA committee decided to develop the software for ADESOS/OCTS data.

The Ocean Color and Temperature Scanner (OCTS) was develop by national space Agency of Japan (NASDA), and it was loaded on ADEOS satellite, and it was an optical radiometer of achieve highly sensitive spectral measurement with 12 bands covering visible and thermal infrared region.

OCTS data is defined by some kind of products, such as local area coverage (LAC) , global area coverage (GAC) and real time coverage (RTC).Moreover, OCTS products are distributed as some standard processing level. For example, LAC is consisted from 3 products of level 1A, level 1B and level 2 and GAC is consisted from 4 products of level 1A, level 2, 3 binned and level 3 binned map.

The fact shows NASDA doesn't always distributed all products required by Scientists. Therefore, a development of a software package is expected to process user specified products.

2. Methodology
The method of radiometric correction algorihm are shown by two parts, because the processing of radiometric correction is completely different visible and near infrared data from thermal infrared data.

The algorithm was developed based on ADEOS Spacecraft Orbital Operations Handbook issue by NASDA.

2.1 Visible and Near Infrared
Band 1 to band 8 of OCTS data are VISIBLE AND NEAR INFRARED SPECTRAL BAND. The radiometric algorithm were shown in next flow.

(1) Input parameter and correlation table
Many kinds of parameter table were prepared by sensor condiguration design, pre-lunch experiments and calibration test.

Global attributes, Scan line attributes, Converted telemetry, Calibration and Sub-sampling table were used to process in correlation. The parameter of the table shows in the follwing.
#Tilt, total gain and offset, temperature of sensor and black body, amp of correlation coefficient, and so on. Also, the radiance and temperature conversion was used that processing time is more faster than calculation to Plank function on program execution.

(2)Principals
The calculation method of gain parameter shows the next equation.

Gain =TL(I, q) x (1 + Bd(Td-Td0) x (1 + Bp(Tp-Tp0) x (1 + Bam(Tam-Tam0) x (1 +Bad(Tad-Tad0)

Where;
I: band number
q: tilt angle
TL : reflectance correlation function by tilt
Bd: temperature correlation function by detector
Td : temperature of detector
Td0: origin temperature of detector
Bp: temperature correlation function by pre-amplifire
Tp: temperature of pre-amplifire
Tpo: origin temperature of pre-amplifire
Bam: temperature of AMUX
Tam0:origin temperature of AMUX
Bad: temperature correlation function by ADC
Tad: temperature of ADC
Tad0: origin temperature of ADC

The calculation method of brightness shows the next equation

Where;
AD:A/D conversion function
VC : OCTS data
Of : total offset
Ga: Total Gain
Gain: gain coefficient by equation 1.

2.2 Thermal Infrared
Band 9 to band 12 of OCTS data are thermal infrared spectral band. The radiometric algorithm is shown in nest step.

Calculation of original black body temperature
Input Brightness of black body
Calculation the average of black body except abnormal value
Input total offset
Calculation of correlation coefficient
Calculation of Brightness
Calculation of albedo

Five black body are loaded on near OCTS senor. Normally, number 1 to 4 of the black body are used. Number 5 black body is compared when number 1 to 4 of the black body output abnormal value. The correlation coefficient is leaded by next equation.

Where:
VB: average of Black body
O: total gain

3.Experiments
The LAC and GAC data of OCTS level 1A data were processed developed algorithm. After processing results were compared from LAC of OCTS level 1B and BGAC of level 2. Naturally, the result of GAC must be compared from level 1B, because there is no level 1B of GAC product.

Conclusions
From the results, the striping noises are gone out and the result were not so different from level 1B product distributed by NASDA.

We noticed from temperature of black body were not stable. Therefore, we must develop new algorithm considered unstable temperature for high accuracy correction.

Finally, the PaNDA committee will distribute the software package via internet. The next address described PaNDA software in detail.

http://www.tric.u-tokai.ac.jp/research1/panda/panda.html

Acknowledgements
The another thanks the members of PaNDA committee who help us to carry out this project. Especially, the author is deeply indebted to prof. Dr. Haruhisa Shimoda who is the chair of the PaNDA committee, for providing valuable comments and advice.

Also, the author thanks Dr. Masao Moriyama with Nagasaki University gave us the table to convert brightness to temperature for OCTS. Finally the author

thanks NaSDA OCTS data for this study.

Reference

NASDA, ADEOS Spacecraft Orbital operations Handbook, Vol. 9 OCTS, 1995
NASDA, OCTS Products Format Description, Version 1.3.1 , 1996.
PaNDA committee, the Manual of PaNDA, 1997.

(1) Original image	(2) Corrected image
Fig 1 Result of Radiometric Correction