Block Clustering Method for Large Scale InSAR Interferometric Phase Diagram based on GSOM
In large-scale complex geographical environments, the interferometric phase diagram regions corresponding to different surface coverage, topography, etc. generally have different phase quality, coherence and signal-to-noise ratio. Segmenting the interferometric phase diagram of interferometric synthetic aperture radar (InSAR) in large-scale complex geographical environments into different sub-images and using different unwrapping methods according to their characteristics is an effective way to avoid the noise diffusion of low-quality phase regions and reduce the overall phase unwrapping accuracy. In order to reasonably segment the interferometric phase diagram, this paper proposes a block clustering method based on the growing self-organizing map (GSOM) network. First, the large-scale interferometric phase diagram is evenly divided into several sub-images of the same size. The correlation coefficient, average phase variance, average pixel value and horizontal and vertical coordinates of the center position corresponding for each sub-phase diagram are calculated, and used as the input feature data of the GSOM network. Then, the square Euclidean distance between the input vector and the weight vectors of all self-organizing map (SOM) nodes is used as the distance metric function. The GSOM network with dynamic growth ability is constructed and trained autonomously to obtain the clustering results of all sub-phase diagrams. The final clustering result of each type is taken as an unwrapped sub-region of the large-scale interferometric phase diagram. To illustrate the effectiveness of the proposed method, SOM network, hierarchical clustering, cuckoo search algorithm are used for comparative experiments with the proposed method, and clustering evaluation indicators are used to evaluate the clustering results. The experimental results show that the proposed method is superior to the above three methods in the clustering of large-scale interferometric phase diagram.
Authors
- Yu Wang (ORCID: https://orcid.org/0000-0003-1563-7306)
- Zhaoxia Wang
Institutions
- Inner Mongolia University of Technology (CN)
Publication Details
- Published
- 2024-09-20
- DOI
- https://doi.org/10.1109/eit63098.2024.10762717
- Primary Topic
- Synthetic Aperture Radar (SAR) Applications and Techniques
- Type
- article
- Field-Weighted Citation Impact
- 0.00
Funders
- Natural Science Foundation of Inner Mongolia
- Inner Mongolia University of Technology