Dear Syarif Budiman,

You can add atmospheric correction to your data preprocessing only if you have a radiative transfer model. Some models are free and available on the web (e.g. 6SV : http://6s.ltdri.org/pages/run6SV.html).

Step 1- You follow the instructions and set the caracteristics of the band you want to correct on 6SV site web.

Step 2- The model will give you result with 2 values: Total Gaseous Transmittance and atmospheric relfectance.

Step 3 - Correct your data using this formula in a GIS software : (Px-Ratm)*TGT

with Px individual values of each pixel in the selected band, Ratm atmospheric relfectance; and TGT total gaseous transmittance.

Step 4 - You do the same for each band.

I hope this helps!

Dear Florent Taureau

Thank you, but I believe in 6S you still need some information related to atmospheric contents, such as aerosol, isn't it?

Aerosol contents are automatically estimated when you choose latitude and select the surronding environment (coastal, urban, continental, ...). You don't need specific data or field measurement.

I am the person who developed the 'Improved DOS' and 'COST' atmospheric corrections models (the COST model is in the paper titled 'Image Based Atmospheric Corrections --- revisited and improved' published in 1996). Keep in mind that either the DOS or COST corrections can be done with a simple stretch function which all software package contain somewhere within their tool boxes of processing procedures.

A very critical aspect of image based correction models, such as these two, is that a 'VALID DARK OBJECT' that can be used to compute the haze values exist within the image. At times the minimum DN value of the image is used blindly as the dark object for atmospheric haze corrections which can lead to over correcting for haze and the surface reflectance values will be under estimated. For information on the Improved-DOS and COST techniques look at the papers published in 1988, 1989, and 1996 (included on the research gate pub list).

Link to slides about Radiometric Calibration and Atmospheric corrections using DOS / Improved DOS / COST models:

https://www.researchgate.net/publication/275042873_Presentation_-_Radiometric_Calibration_and_Atmospheric_Corrections

Link to COST paper:

https://www.researchgate.net/publication/236769129_Image-Based_Atmospheric_Corrections_-_Revisited_and_Improved

Link to the Improved DOS paper:

https://www.researchgate.net/publication/223795843_Chavez_PS._An_improved_dark-object_subtraction_technique_for_atmospheric_scattering_correction_of_multispectral_data._Remote_Sensing_of_Environment

Data Presentation - Radiometric Calibration and Atmospheric Corrections

Article Image-Based Atmospheric Corrections - Revisited and Improved

Article An improved dark-object subtraction technique for atmospheri...

Dear Florent and Pat

Thank you very much for your explanation.

And Pat thank you for the link to your papers. Indeed to choose a valid dark object sometimes quite difficult.

Thank you to both of you for your answer and explanation.

Hi,

Some addition to what have been discussed so far. 

Broadly, atmospheric correction is a two step process:

1. simulate atmospheric transmittance: you can use radiative transfer code for this (6S, MODTRAN)

2. From the simulated transmittance derive correction parameters, one such approach is given in (De Haan and Kokke, 1996; De Haan et al., 1997). 

Apply the correction parameter to each pixel. Look into this paper where a brief of the method is given pg 6-7.

Compliments to Nitin for good reply.