ATCOR - Code Comparison

The table below gives an overview of the features of the ATCOR packages as distributed through ReSe (ATCOR-IDL) in comparison to current alternative codes, as available on the market:

System Features        
supported operating systems Windows, Linux, MacOSX Windows, Linux Windows, Linux Windows, Linux
base software none 1) Geomatica or Erdas Imagine none ENVI
airborne yes (A-4) no yes yes
satellite yes (A-23) yes yes yes
hyperspectral yes 3) limited yes yes
ultraspectral (<2 nm) 2) limited 3) limited 3) yes yes
extreme tilt angles 2) no no yes yes
batch processing yes yes yes no
image tiling yes yes no yes

Correction Features

variable visibility yes yes yes yes
aerosol type detection yes no no no
adjacency effect yes yes no yes
water vapor retrieval yes 4) no yes yes
haze removal yes yes no no
cirrus cloud removal yes no no no
cloud shadow removal yes no no no
preview of spectra yes yes yes no
inflight calibration yes (one target) no no
normalizing to nadir view (wide FOV imagery) yes no no no
Physical Observer BRDF correction yes no no no
spectral polishing yes 5) limited no yes 5)
spectral calibration yes no yes yes
spectral smile correction yes no yes no
empirical line correction no no yes yes (ENVI)

Thermal region

surface temperature yes yes no yes (ENVI)
surface emissivity yes yes no yes (ENVI)

Rugged terrain

terrain height (DEM) yes yes no no
DEM illumination effects yes yes no no
empirical BRDF illumination correction yes yes no no

Notes and Remarks:

1) ATCOR from ReSe runs on the free IDL Virtual Machine.

2) It's a true advantage of other atmospheric correction softwares (e.g. FLAASH or ACORN) that sensor-specific LUTs can be calculated directly through Modtran®. This feature is particularly well suited for sensors at tilted angular conditions with large tilt angles. Also, 'ultraspectral' sensors having resolutions below 2nm are not ideally sampled in the ATCOR LUT. On the other hand, our experience has shown that 99% of all currently relevant data acquisitions are within the range which is covered by the ATCOR pre-calculated LUTs.

Note: The ATCOR LUTs are based on the accurate scaled DISORT (8 streams) multiple scattering option in atmospheric window regions and the accurate correlated-k algorithm in absorption regions. A variable spectral wavenumber resolution in MODTRAN is employed to achieve a constant wavelength resolution of 0.4 nm in the LUTs.

3) The IDL version of ATCOR supports the inclusion of new (user defined) sensors. A resampling tool contained within ATCOR allows interpolation of virtually any multispectral or hyperspectral sensor from the database (at 0.4 nm resolution). Inaccuracies due to this interpolation become only relevant at spectral resolutions below 2 nm.

4)   ATCOR uses the latest release of Modtran®5 (v2r1) and its LUTs are based on the HITRAN atlas of 2007 - a version which to our knowledge corrects earlier problems in the water vapor absorption bands. We will include any newer version of Modtran® as soon as it is available.

5) The spectral polishing features within FLAASH and ENVI are very sophisticated. ATCOR also offers such a feature which is about to be improved.

An independent comparison was published by Eyal Ben-Dor (2005) and indicates that the performance of ATCOR compares very well with other products available on the market, see: E. Ben-Dor, B. C. Kindel, and K. P. K1, "A comparison between six model-based methods to retrieve surface reflectance and water vapor content from hyperspectral data: A case study using synthetic AVIRIS data," presented at International Conference on Optics & Optoelectronics - ICOL 2005, Dehradun, India, 2005.

As ATCOR is continuously improved, we would be happy to participate in independent validation exercises and we would offer free time-limited licenses for such purposes.

Home | News | Software | Support | Download | Company

last modified: DS, 04.04.16