SBATool

Split-Based Approach (SBA) for SAR-based Change Detection. This tool contains two algorithms:

• GSBA: Growing Split-Based Approach. Parallel computation is supported.
• HSBA: Hierarchical Split-Based Approach.

For detailed information about HSBA, refer to:

Chini, M., Hostache, R., Giustarini, L., & Matgen, P. (2017). A hierarchical split-based approach for parametric thresholding of SAR images: Flood inundation as a test case. IEEE Transactions On Geoscience and Remote Sensing, 55(12), 6975-6988. https://doi.org/10.1109/TGRS.2017.2737664

Detailed information about GSBA is still under review (as of 2 Oct 2023).

Contents

1. Environment Preparation
2. Input Data Files
3. Input Config Files
4. Running Growing Split-Based Approach (GSBA)
5. Running Hierarchical Split-Based Approach (HSBA)
6. Plot and Fit Histogram (for data inspection)
7. Reference

---

Environment Preparation

It is recommended that you run SBATool in MATLAB 2019b or later versions. Note: Errors may appear if you run in MATLAB 2023 or later versions.

Add SBATool to MATLAB path by doing:

addpath(genpath('{SBATool folder}'))  %{SBATool folder} is the path to SBATool

You will also need to install the following MATLAB Toolbox:

Deep Learning Toolbox

Input Data Files

An example of the input files are shown in the example folder:

File Name	Description	Required/Optional
lumberton.tif	Z-score map	Required
lumberton_mask.tif	Layover and shadow mask 0=non-masked other values=masekd	Optional
lumberton_hand.tif	HANDEM values in meters	Optional
lumberton_lia.tif	Local incidence angles (LIA) in degrees	Optional
lumberton_val.tif	Validation data 0=no change 1=change This dataset is available only upon request	Optional

All input files need to share the same prefix. The suffix can be .tif (recommended), .img or any isce suffix.

If any of the optional files is not present in the folder, the corresponding step (masking/hand masking/lia masking/validation) will be omitted.

The first four tif files need to be in the same image size. The validation file can have a different image size, usually smaller than the other four files.

All files need to be in the same reference frame.

For information about HANDEM, refer to:

Rennó, C. D., Nobre, A. D., Cuartas, L. A., Soares, J. V., Hodnett, M. G., Tomasella, J., & Waterloo, M. J. (2008). HAND, a new terrain descriptor using SRTM-DEM: Mapping terra-firme rainforest environments in Amazonia. Remote Sensing of Environment, 112(9), 3469-3481. https://doi.org/https://doi.org/10.1016/j.rse.2008.03.018

For information about LIA, refer to:

Shibayama, T., Yamaguchi, Y., & Yamada, H. (2015). Polarimetric Scattering Properties of Landslides in Forested Areas and the Dependence on the Local Incidence Angle. Remote Sensing, 7(11). https://doi.org/10.3390/rs71115424

Input Config Files

Example of the config files are shown in the config folder:

File Name	Description	Required/Optional
config_flood_gsba.txt	Job configuration file for GSBA	Required
config_flood_hsba.txt	Job configuration file for HSBA	Required

You should run different algorithms under different folders. For example, to test the gsba example, do:

cp example example_gsba -rf
cp config/config_flood_gsba.txt example_gsba

The same applies to the hsba example.

Reference

For more information about the methodology of GSBA and SBATool, refer to:

Lin, N. Y., Yun, S.-H., Bhardwaj, A., & Hill, M. E. (2019). Urban Flood Detection with Sentinel-1 Multi-Temporal Synthetic Aperture Radar (SAR) Observations in a Bayesian Framework: A Case Study for Hurricane Matthew. Remote Sensing, 11(15), 1-22. https://doi.org/10.3390/rs11151778

Lin, Y. N., Chen, Y.-C., Kuo, Y.-T., & Chao, W.-A. (2022). Performance Study of Landslide Detection Using Multi-Temporal SAR Images. Remote Sensing, 14(10). https://doi.org/10.3390/rs14102444