Evaluating the Cloude-Pottier and Freeman- Durden scattering decompositions for distinguishing between unharvested and post-harvest agricultural fields
Adams, J.R., Rowlandson, T., McKeown, S.J., Berg, A.A., McNairn, H., and Sweeney, S.J. (2013). "Evaluating the Cloude-Pottier and Freeman- Durden scattering decompositions for distinguishing between unharvested and post-harvest agricultural fields", Canadian Journal of Remote Sensing, 39(4), pp. 318-327. doi : 10.5589/m13-040 Access to full text
This study evaluates the utility of the Cloude–Pottier and Freeman–Durden scattering decompositions for providing agricultural land surface information during autumn months using C-band polarimetric RADARSAT-2 data. We applied these decompositions over 94 agricultural fields in Southern Ontario, Canada, to characterize scattering mechanisms from unharvested senesced crops and harvested fields with three generalized soil tillage practices. The decompositions were applied to RADARSAT-2 images over six dates during October and November 2010 at high (49°) and low (23°) incidence angles. Agreement was found between the decompositions for the identification of primary (volume and rough surface scatter) scattering mechanisms for the senesced unharvested crops and the harvested fields. Significant statistical separability was observed between the strengths of decomposition parameters when comparing (i) senesced unharvested crops to post-harvest conventional tillage fields and (ii) post-harvest no tillage fields to post-harvest conventional tillage fields. These results suggest that high accuracy classifications may be possible with these data; however, weak separability was observed when comparing fields with conservation tillage. The strongest separability was observed with Entropy and α-angle of the Cloude–Pottier decomposition and the rough surface scattering component of the Freeman–Durden decomposition, suggesting sensitivity of these parameters to surface roughness and crop residue. Results also demonstrated that superior separability was found with the data at the higher 49° incidence angle in contrast to data acquired at the lower 23° incidence angle imagery.
- Date modified: