SFE2D-Demo

Demo version of SFE2D: a 2D tool for spatial and spectral feature extraction from geo-images (PCA, ICA, CWT, S-PCA, S-ICA). The demo is distributed in P-code form (.p files) with SFE2D.fig. Try the workflow with limited parameters; for full functionality, use the Full version.

Lead Developer: Bahman Abbassi
Principal Investigator: Li-Zhen Cheng
Affiliation: Institut de Recherche en Mines et en Environnement (IRME), Université du Québec en Abitibi-Témiscamingue (UQAT)

---

What this demo includes

• Same GUI and workflow as SFE2D: load imagery and/or point data, run spatial (PCA, kICA, nICA) and spectral (CWT, S-PCA, S-kICA, S-nICA) extraction, visualize and export.
• Demo limits (full version removes these):
  • Inputs: 1 satellite image, 1 XYZ (point) dataset.
  • Grid: max 500 pixels in X and Y (suggested spacing shown if exceeded).
  • Spatial ICA: convergence fixed at 1e-4, max iterations 20.
  • Spectral: max 3 scales, 6 angles; mother wavelet Gaussian only; S-ICA convergence 1e-4, max iterations 20; CWT Filter and Scale Dilation fixed at 1.
  • Segmentation: max 6 colors.

When a limit is reached, the GUI shows a message and directs you to the full version.

---

Requirements

• MATLAB: R2021a (9.10) or later; tested up to R2025b.
• Toolbox: Wavelet Toolbox (for 2D CWT).
• System: Windows 7/8/10; ≥ 8 GB RAM recommended.

---

Installation & run

1. Clone or download this repository.
2. Add the folder containing the P-codes (.p files) and SFE2D.fig to your MATLAB path, or set it as the current directory.
3. Ensure SFE2D.fig is in the same folder as SFE2D.p.
4. In the MATLAB Command Window:

   SFE2D

---

Workflow (short)

1. Preprocessing — Set bounds (Min/Max Lon/Lat), Spacing (Arc-Sec) (use suggested value if you hit the 500-pixel limit), load one Image and/or one XYZ Data CSV.
2. Spatial — Choose Select Inputs (Satellite / Point Data / All), then run PCA, kICA, and/or nICA.
3. Spectral — Set CWT Inputs, scales (≤3), angles (≤6), mother wavelet Gaussian, then CWT, then S-PCA, S-kICA, and/or S-nICA.
4. Plot & export — Use Plot Results and Exporting (Surfer grids, XYZ CSV).

---

Repository contents

The demo is provided as P-codes (.p files) plus the GUI layout file:

• SFE2D.p, SFE2D.fig — Main GUI (P-code; demo limits enforced).
• PCA.p, fastICA.p, fastICA_CWT.p, kmeans_fast_Color.p, fig.p, FilterB.p, grd_write.p, centerRows.p, whitenRows.p, PlaceThresholdBars.p — Core routines (P-code).

---

Full version

For unrestricted parameters, multiple inputs, and all wavelets, use the Full version of SFE2D:

https://github.com/bahmanabbassi/SFE2D-3.0

---

References

• Abbassi, B., Cheng, L.Z. SFE2D: A Hybrid Tool for Spatial and Spectral Feature Extraction. In: Mining Technology. IntechOpen, 2022. DOI: 10.5772/intechopen.101363
• Abbassi, B., Cheng, L-Z., 2020, SFE2D : Un outil d'extraction de caractéristiques spectrales de géo-images. Rapport pour le ministère de l'Énergie et des Ressources naturelles, Bureau de la connaissance géoscientifique du Québec. Lien (SIGEOM)
• Abbassi, B., and Cheng, L-Z., 2019, Integrated feature extraction in high-dimensional geophysical imaging. Présentation à Québec Mines + Énergie 2019. DOI: 10.13140/RG.2.2.36577.93286
• SFE2D User Manual (ver. 2.0), April 2021.

Dependencies

• Fast K-means Algorithm for clustering Color Image: ankit dixit (2026). Fast kmeans Algorithm Code, MATLAB Central File Exchange. Retrieved March 16, 2026. Link
• Fast-ICA and PCA: Brian Moore (2026). PCA and ICA Package, MATLAB Central File Exchange. Consulté le 16 mars 2026. Lien

---

License & copyright

© B. Abbassi, 2022. All rights reserved. This demo is for evaluation only. For licensing and the full version, see the author’s GitHub and User Manual.