Abstract:In this study, multifrequency, Polarimetric Synthetic Aperture Radar (PolSAR) images are employed to test the performance of a novel fully polarimetric speckle filter based on Morphological operations. PolSAR images of the P, L, & C microwave bands from the AIRSAR satellite are used. The focus is to determine the frequency at which the filter performance is the best. Firstly, multi-looking is performed on the PolSAR data to enhance radiometric resolution. The multi-looking operation also provides higher sensitivity to brightness changes and less noise. The scattering [S] matrix is converted into the coherency [T] matrix and the proposed filter is individually applied on each of the 9 elements of the [T] matrix. Two successive opening followed by closing operations are applied to test the filter for idempotency. The filtered images are then inspected for photographic quality followed by a classification study. The visual quality of the filter is determined using different full references as well as no reference quality metrics. After the visual inspection the filtered images are classified using Support Vector Machine (SVM), with Radial Basis Function as the kernel, along with cross validation. A comparative study of the proposed filter with existing filters is carried out using quality indices and classification accuracies. For comparison purpose the P, L, & C Band PolSAR images of the same location are used. The overall classification accuracies at P, L & C band frequencies are found to be 95.48%, 99.85% & 94.81% respectively. Since the L-Band image gives the best classification accuracies, PolSAR images from the L-Band ALOS-1 satellite of the San Francisco region, USA and ALOS-2 Data of the Vietnam region are used to inspect the filter performance using 8 different classifiers. It is observed that the fully polarimetric filter based on morphological operations gives the best results on L-Band data. Also, the classification accuracies of all the classifiers are high after applying the proposed filtering technique. The proposed filter can reduce the speckle significantly, enhance point targets and maintain visual quality.

Speckle filtering of polarimetric SAR image and the enhancement for classification

Speckle filtering in polarimetric SAR data based on the subspace decomposition

Speckle Filtering Method Based on Polarimetric SAR Image

A PolSAR Image Speckle Filtering Method Preserving Point Targets and Dominant Scattering Mechanisms

Adaptive Speckle Filter for Multi-Temporal PolSAR Image with Multi-Dimensional Information Fusion

Feature-Based Nonlocal Polarimetric Sar Filtering

New Method of Feature Extraction in Polarimetric Sar Image Classification

Nonlocal Filtering for Polarimetric SAR Data: A Pretest Approach

Classification of Full-Polarization SAR Images Combined with ICA-Based Speckle Reduction

Superpixel-Based Classification Using K Distribution And Spatial Context For Polarimetric Sar Images

Polarimetric SAR Speckle Filtering Based on Similarity Test and Adaptive Clustering

A Multi-channel Extension of Polarimetric Whitening Filter

From Pixel to Superpixel: A Multi-Scale Strategy for Polarimetric Sar Image Classification

Edge Detection In Polarimetric Sar Images Based On The Nonsubsampled Contourlet Transform

Multi-Frequency Analysis of Fully Polarimetric Speckle Filter Based on Image Morphology for Land Use Classification

Polsar Image Classification Based on Three-Dimensional Wavelet Texture Features and Markov Random Field.

Polarimetric Whitening Filter for POLSAR Image Based on Subspace Decomposition

Fully Polarimetric SAR Image Filtering by Combining Block Matching and Lee Filter

Superpixel-based classification of polarimetric synthetic aperture radar images

Deep Learning Based Speckle Filtering for Polarimetric SAR Images. Application to Sentinel-1

Linear Feature Detection in Polarimetric Sar Images