This study aims to determine the best optimization function for a feedforward neural network using fully polarimetric Quad-Pol data from the ALOS-2 (L-Band) satellite. Data from the Mumbai Region, India, is used to evaluate the performance of different optimization functions. Twelve different functions are considered, including Levenberg-Marquardt, Bayesian Regularization, BFGS Quasi-Newton, Resilient Backpropagation, Scaled Conjugate Gradient, Conjugate Gradient with Powell/Beale Restarts, Fletcher-Powell Conjugate Gradient, Polak-Ribiére Conjugate Gradient, One Step Secant, Variable Learning Rate Gradient Descent, Gradient Descent with Momentum, and Gradient Descent. The ALOS-2 satellite image is filtered using the Lee Sigma Filter and decomposed into seven components using the Gulab 7 Component decomposition (7SD) technique. Neural Network classifiers with 12 different optimization functions are applied on the decomposed image using 7SD. The trained models are analyzed using scores such as individual class accuracies, Kappa Score, Micro & Macro F1 Score, Micro & Macro Precision, Micro & Macro Recall, and Overall Accuracy. The time required to train the different neural networks is also noted. It is observed that Levenberg-Marquardt performs the best as compared to others

Mangroves are vital coastal ecosystems that provide numerous ecological and socio-economic benefits. Accurate detection and monitoring of mangroves are crucial for effective conservation and management. This work aims to determine the best optimization function for a feed-forward neural network that can be used for Mangrove detection. Fully polarimetric Quad-Pol data from the ALOS-2 (L-Band) satellite is used to train the Neural Network. PolSAR data of the Mumbai Region, India is used since it has a huge Mangrove cover. Firstly, the PolSAR image is filtered using the most used Lee Sigma Filter. The filtered image is then decomposed using the Gulab seven component decomposition (7SD) technique. These seven components which act as the feature vectors are used to train the Neural Network classifier. While training 12 different optimization functions are used. The trained models are analyzed using the different evaluation metrics as well as the time required to train the neural networks. It is observed that the Neural Network optimized with Bayesian Regularization and Levenberg-Marquardt optimization functions, gives the highest accuracy.

In recent times, Polarimetric Synthetic Aperture Radar (PolSAR) data is available free of cost due to missions like UAVSAR and Sentinel. As ample data is available the applications are infinite. Till today many researchers have developed various techniques to classify PolSAR data efficiently. They have proposed classification techniques for which the ground truth should be available to train the classifier and validate the results for a particular geographic area. Training the classifier for each area is a time-consuming task and hence there is a need to develop a generalized model which can classify any geographical area acquired from a specific sensor for various land-cover features like water, settlement, forest, wetlands etc. In this paper a generalized machine learning model is proposed which can classify the data acquired from ALOS-PALSAR-2 L-band, irrespective of geographical area for the same land cover features. ANN classifier is used in this work. The classifier is trained using Mumbai and tested for San Francisco and Delhi data. It is observed that the classification accuracy for San Francisco as well as New Delhi is high.

Urban area classification is one of the important applications of Polarimetric Synthetic Aperture Radar (POLSAR). This paper suggests an effective technique for classifying the terrain using different polarimetric decompositions and circular correlation coefficient (CCC) along with the total power. It is observed that after applying decomposition, the oriented urban area is getting classified as forest. The result of proposed technique shows that the classification accuracy increases significantly for settlement class i.e., most of the oriented settlement gets classified as settlement not as forest.

For more than four decades, it has been known that the dielectric constant of the lunar surface can be retrieved from the Fresnel reflection coefficients. However, theoretical models have met with limited success in validating laboratory test results from the Apollo missions to date. This paper is the first study to focus on the use of high-resolution full-polarimetric synthetic aperture radar datasets for the retrieval of the dielectric constant of the lunar surface from the Fresnel reflection coefficients. We initially show that it is possible to retrieve the lunar dielectric constant via the classical Freeman-Durden Decomposition (FDD). The performance of the FDD algorithm is found to be unacceptable over regions with surface slopes and craters, and for sub-surface soil samples. Accurate estimation is not possible by simply replacing the volume scattering model in the FDD with popular and widely used volume scattering models. Therefore, a model-based three-component decomposition (TCD) algorithm for a robust retrieval of the lunar dielectric constant is proposed. The proposed TCD algorithm implements an efficient branching condition combined with double unitary matrix rotations and provides exceptionally accurate dielectric constant estimation. The proposed TCD algorithm is validated by using L band full-polarimetric datasets acquired by the Chandrayaan-2 mission over Apollo 12, Apollo 15, and Apollo 17 landing sites. Comparisons are also made with other three-component decomposition algorithms. Excellent agreement between the estimated values by the proposed TCD algorithm and the reference values for the dielectric constant, available from the literature, has been observed.

A new dielectric constant estimation approach for the lunar surface using Mini-RF synthetic aperture radar (SAR) data and Chandrayaan-2 SAR data is presented in this work. Both the SAR systems are based on a hybrid-polarimetry (hybrid-pol) configuration that transmits a circularly polarized wave and coherently measures the backscattered wave by dual orthogonal linearly polarized channels. From the three-component hybrid-pol SAR decomposition technique, the ratio of the Fresnel reflection coefficients for horizontal and vertical polarization transmission can be estimated. This ratio, which is referred to as the co-polarization ratio, is used by the proposed methodology to find the real value of the dielectric constant. For performance validation, the proposed method is implemented on the Mini-RF hybrid-pol SAR data acquired over Apollo 17 landing sites. The values of the real part of dielectric constant are estimated for six different regions covering the collection sites of six Apollo 17 samples: 72 441, 73 241, 74 241, 75 081, 76 001, and 79 135. The results obtained using the proposed method are found to be in good agreement with the laboratory-measured results of the corresponding samples. Furthermore, the proposed methodology is also implemented on the Chandrayaan-2 hybrid-pol SAR data acquired over theBiot craterregion. Various small areas possessing different possible surface characteristics, situated inside and outside theBiot crater, are being analyzed for validation.

RISAT-1 data is acquired over Mumbai in hybrid and linear dual polarizations. The mean and standard deviation of backscattering coefficients ( σ0) are computed and have been analyzed for various land features. Classification accuracy between RISAT-1 hybrid and dual polarimetric data has been compared. The effect of different multilook on the classification accuracy is also studied. Wishart supervised and Support Vector Machine (SVM) classifiers are used for this study. It has been observed that the classification accuracy can be improved by using m-δ or m-χ decomposition along with Circular Polarization Ratio (CPR) and SPAN of hybrid polarimetric data.

The target decomposition techniques give more information about scattering mechanism than obtained through covariance or coherency matrix. In this paper, the effect of various parameters of Touzi decomposition on classification accuracy is studied. The work shows that out of many Touzi parameters, the first components of α, Φ, λ, τ along with span can effectively classify various land features. The influence of helicity is more prevalent at L-band compared to C-band. The effect of different non-parametric classifiers on classification accuracy is also studied.

Speckle has a nature of multiplicative noise which is difficult to deal as compared to additive noise. It complicates the problem of interpretation of the image segmentation and classification. The primary goal of existing speckle filtering algorithms, which are subjective in nature is to reduce the speckle without loss of information. Various techniques have been proposed to suppress the speckle. In this paper we propose Self-Organizing Feature Map (SOFM) based polarimetric SAR speckle filter. The filter is evaluated using fully polarimetric ALOSPALSAR and Radarsat-2 data imaged over Mumbai, India. Quantitative and qualitative results revels that SOFM based approach is effective in terms of bias and speckle reduction.

The launch of RISAT-1 Indian remote sensing satellite on 26thApril 2012, made it possible to collect hybrid polarimetric data from a space-borne sensor. The RISAT-1 C-band compact polarimetry data acquired over Mumbai is analyzed and assessed for classification of various land features and also compared with other fully polarimetric spaceborene SAR data sets. For better comparison, RISAT-1 C-band and RADARSAT-2 C-band simulated compact polarimetric data is classified and compared.