Image Fusion Based on NSCT
- Login to Download
- 1 Credits
Resource Overview
A ready-to-run remote sensing image fusion program based on Non-Subsampled Contourlet Transform (NSCT). This advanced transform, evolved from the Contourlet transform, features multi-scale and multi-directional decomposition capabilities while employing a non-subsampling approach that eliminates shift variance between subband images. The implementation involves image decomposition through NSCT, application of appropriate fusion rules, and final reconstruction via inverse NSCT transformation to produce fused images.
Detailed Documentation
This NSCT-based remote sensing image fusion program can be executed directly without additional configuration. The core algorithm utilizes Non-Subsampled Contourlet Transform (NSCT), an enhancement of the standard Contourlet transform that provides superior multi-scale and multi-directional analysis. The non-subsampling methodology preserves spatial consistency across all subbands, eliminating shift variance issues common in traditional transforms. The implementation workflow includes: decomposing input images using NSCT filters, applying pixel-based or region-based fusion rules to combine coefficient maps, and performing inverse NSCT reconstruction. Key functions likely involve nsctdec() for decomposition, fusion rule functions for coefficient processing, and nsctrec() for reconstruction, ultimately generating high-quality fused images with preserved spectral and spatial characteristics.
- Login to Download
- 1 Credits