+- ReSe User Support Forum (https://forum.rese-apps.com)
+-- Forum: Support Discussions (https://forum.rese-apps.com/forumdisplay.php?fid=9)
+--- Forum: ATCOR-4 (https://forum.rese-apps.com/forumdisplay.php?fid=12)
+--- Thread: Correct hyperspectral imagery of a vertical rock outcrop (/showthread.php?tid=67)
Correct hyperspectral imagery of a vertical rock outcrop - rese_support - 27-Jan-2015
Quote:I'm interested in using ATCOR to correct hyperspectral imagery of a vertical rock outcrop captured with a tripod mounted camera. Radiometric calibration of these images is usually accomplished by viewing one or more reflectance standards within the scene and applying an Empirical Line calibration. However, we would like to move away from the requirement for in situ targets. Is it possible to use ATCOR accomplish the calibration if the camera provides radiance values?
We are also scanning the outcrop with a lidar sensor, so we have the necessary geometric information. However, both the 3D point data and the hyperspectral imagery is captured from a horizontal, rather than vertical, perspective. From what I've read, it appears ATCOR is designed to work with nadir, or nearly nadir, looking cameras.
You understood correctly that ATCOR by design asks for a nadir (2.5 D) data set for its corrections and also it's atmospheric LUTs as derived
from Modtran are 2.5 D only. In order to make it work for the mine-face situation, quite some work would need to be done.
Questions are:
- Which are the distances from the rock face; i.e. could the correction be done as an irradiance correction only or would you need to
include scattering/transmittance effect on the second path?
- Is it possible to bring the LIdar data and the image data in a 1:1 (pixel-wise) relation such that the relative angles of the rocks towards
the sun could be calculated?
- Is the geometric accuracy of the camera good enough for that?
- Is the radiometry of the camera on the tripod accurate enough?
The major potential advantage of an ATCOR-like correction would be the possibility of illumination and shadow correction which can not be
achieved by in-field targets.
If you got a well suitable test case (including the geometry) we could possibly give it a try later this year (after spring time) - my plans
are anyhow to focus on shade corrections this year as there are many open issues in this field.