New change detection technique using ASTER and CBERS-2 images to monitor Amazon tropical forest

Abstract

With the availability of CCD/CBERS-2 images, the objective of this research is to show the technique on "Radiometric rotation controlled by a no-change axis (RCEN)", using a multi-temporal comparison with the image from ASTER/TERRA. The result of this technique is a change-detection image derived from the date integration of both sensors mentioned, whose continuous image (floating-point numbers) was transformed into a thematic image through the procedure of slicing and labelling with four thresholds. This allows the stratification of spectral attributes in five thematic classes: two related to degradation, two referring to regeneration and one of no-change. From this point on a map of changes 2003/2004 was generated, illustrating the distribution of the five thematic categories mentioned at the transition forest zone, which is suffering strongly under the impact of the agricultural frontier, cattle raising and also, timber exploitation activities. © 2005 IEEE.