Ant Colony Optimization Based Anisotropic Diffusion Approach for Despeckling of SAR Images

• Published in: Integrated Uncertainty in Knowledge Modelling and Decision Making Vol. 9978; pp. 389 - 396
• Main Authors: Bhateja, Vikrant, Tripathi, Abhishek, Sharma, Aditi, Le, Bao Nguyen, Satapathy, Suresh Chandra, Nguyen, Gia Nhu, Le, Dac-Nhuong
• Format: Book Chapter
• Language: English
• Published: Switzerland Springer International Publishing AG 2016; Springer International Publishing
• Series: Lecture Notes in Computer Science
• Subjects: Anisotropic Diffusion (AD); Ant Colony Optimization (ACO); Artificial intelligence; Data mining; Despeckling; Information retrieval; Iterations
• ISBN: 9783319490458; 3319490451
• ISSN: 0302-9743; 1611-3349
• DOI: 10.1007/978-3-319-49046-5_33

Synthetic Aperture Radar (SAR) images are known to be corrupted by granular noise known as speckle. This noise is inherently present in these images owing to acquisition constraints and is a major cause of visual quality degradation. The anisotropic diffusion approaches for despeckling are constrained in terms exercising control over the non-homogeneous regions. This paper proposes to improve the non-linear Anisotropic Diffusion (AD) filter for despeckling using Ant Colony Optimization (ACO) algorithm. The main essence of this work is to suppress speckle and preserve the structural content. The issue of residual speckle content has been minimized by optimal selection of AD parameter(s) using ACO algorithm. Experimental results advocate the performance improvement achieved and has been validated using objective measures of image quality evaluation.