| Paper Details: | Downloads: 2012 |
| Serial Number: | P1151514367
|
| Title: | Mutual Use of Derived Information from Radarsat-2 and ETM-8 Data for the Improvements of Feature Extraction
|
| Authors: | Ashraf K. Helmy and Ayman H. Nasr and Safaa M. Hassan
|
| Abstract: | A new approach to improve the feature extraction and map updating of part of Halayib area located in south eastern desert of Egypt, using multisource data sets, is proposed in this paper. It investigates the potential of the joint use of derived information from both PolSAR and Optical images. As they refer to different properties of the observed scene, it’s expected to complement each other. The approach is consisted of combining Radarsat-2 Pauli decomposition data, with ETM-8 principal component transformation. We used the bandlet transform to decompose the original images into a series of frequencies. It depends on the geometric flow of an image to construct the bandlet basis. The geometric flow takes advantage of the geometric regularity of the image structure and represents sharp image transitions such as edges efficiently in the output image. For producing the fused image, we proposed a fusion rule to first-rate the geometric flow of the source images, and then employed the m-PCNN to implement the fusion process with the bandlet coefficients. The new results were evaluated using qualitative and quantitative measures. They demonstrate improved interpretation, where PolSAR texture information enhances the fusion product. The proposed approach clearly states that, the combination of PolSAR and Optical derived data has potential to improve feature extraction and map updating in desert area.
|
| Keywords: | PolSAR, RADARSAT-2, ETM-8, Data Fusion, Bandlet transform, PCNN
|
| Journal/Conference: | International Journal of Graphics, Vision and Image Processing
|
| Volume: | 15
|
| Issue: | 1
|
| Submission Date: | 3/28/2015 12:00:00 AM
|
| Review Date: | 4/23/2015 12:00:00 AM
|
| Publishing Date: | 5/10/2015 12:00:00 AM
|
| Article Downloads: | 2012
|
| Download: |
|