Motofumi Arii

Model-Based Decomposition of Polarimetric SAR Covariance Matrices Constrained for Nonnegative Eigenvalues

Model-based decomposition of polarimetric radar covariance matrices holds the promise that specific scattering mechanisms can be isolated for further quantitative analysis. In this paper, we show that current algorithms suffer from a fatal flaw in that some of the scattering components result in negative powers. We propose a simple modification that ensures that all covariance matrices in the decomposition will have nonnegative eigenvalues. We further combine our nonnegative eigenvalue decomposition with eigenvector decomposition to remove additional assumptions that have to be made before the current algorithms can be used to estimate all the scattering components. Our results are illustrated using Airborne Synthetic Aperture Radar data and show that current algorithms typically overestimate the canopy scattering contribution by 10%-20%.

Adaptive Model-Based Decomposition of Polarimetric SAR Covariance Matrices

Previous model-based decomposition techniques are applicable to a limited range of vegetation types because of their specific assumptions about the volume scattering component. Furthermore, most of these techniques use the same model, or just a few models, to characterize the volume scattering component in the decomposition for all pixels in an image. In this paper, we extend the model-based decomposition idea by creating an adaptive model-based decomposition technique, allowing us to estimate both the mean orientation angle and a degree of randomness for the canopy scattering for each pixel in an image. No scattering reflection symmetry assumption is required to determine the volume contribution. We examined the usefulness of the proposed decomposition technique by decomposing the covariance matrix using the National Aeronautics and Space Administration/Jet Propulsion Laboratory Airborne Synthetic Aperture Radar data at the C-, L-, and P-bands. The randomness and mean orientation angle maps generated using our adaptive decomposition significantly improve the physical interpretation of the scattering observed at the three different frequencies.

A General Characterization for Polarimetric Scattering From Vegetation Canopies

Current polarimetric model-based decomposition techniques are limited to specific types of vegetation because of their assumptions about the volume scattering component. In this paper, we propose a generalized probability density function based on the nth power of a cosine-squared function. This distribution is completely characterized by two parameters; a mean orientation angle and the power of the cosine-squared function. We show that the underlying randomness of the distribution is only a function of the power of the cosine-squared function. We then derive the average covariance matrix for various different elementary scatterers showing that the result has a very simple analytical form suitable for use in model-based decomposition schemes.

Efficient Motion Compensation of a Moving Object on SAR Imagery Based on Velocity Correlation Function

Current motion compensation algorithms for a moving object on synthetic aperture radar (SAR) imagery by refocusing are accurate, but they suffer from a severe computational efficiency problem. If the step size of an assumed target velocity is too large, one could miss the true target velocity, whereas too small a step size would lead to a considerable computation load. To address this issue, the way motion error affects a compressed signal should be understood. In this paper, we introduce the concept of velocity correlation function (VCF) to describe the sensitivity of an SAR compressed signal to motion error. We propose its analytical model by straightforwardly tackling conventional matched filtering, and then we thoroughly compare the model with the experimentally obtained VCF for each static and moving object on real L-band airborne/spaceborne SAR imagery. The analytically derived VCF is well validated, as it shows excellent agreement with those experiments. Finally, the VCF is generalized to a velocity correlation map to visualize a sensitivity of SAR system parameters to target motion in both the range and the azimuth directions simultaneously.

Improvement of ship-sea clutter ratio of SAR imagery using standard deviation filter

The needs for SAR remote sensing has been rapidly increased to efficiently promote ocean security. In this context, solid ship detection algorithm plays a vital role to execute expected performance. Most of conventional algorithms utilize a Constant False Alarm Rate (CFAR) technique, and have achieved great success. However, there has been no solid procedure to select an appropriate CFAR threshold for various area. In this paper, based on physical scattering theory we proposed a brand-new Standard Deviation (STD) filter by which a dependency to CFAR threshold can be significantly decreased. Validation of the technique was also conducted using Pi-SAR image provided by Japan Aerospace Exploration Agency (JAXA).

Retriveval of soll moisture under vegetation using Polarimetric Scattering Cubes

Soil moisture inversion from polarimetric SAR data has attracted significant attention for the past twenty years. Comparing with the simple case of bare surface, it is extremely complicated for vegetated terrain to invert soil moisture because of a larger number of scattering mechanisms that contributes to the observation. In this paper, we show how polarimetric decomposition technique, which decomposes SAR observations into preferred scattering mechanisms, can be used for the inversion. The result leads us to give up the use of polarimetric decomposition because of an unknown attenuation ratio caused by the canopy. Then a new inversion algorithm using Polarimetric Scattering Cubes (PSC) is introduced with simulation results to show a sensitivity to physical parameter such as vegetation distribution. Finally, we also discuss how the technique should be implemented for the real SAR data.

Sensitivity study of radar backscatter from boreal forest using discrete scatterer model

Polarimetric SAR decomposition has been expected to be a potential candidate of a tool to measure physical parameters of vegetated fields. For this context, it is important to validate theoretical model quantitatively with well controlled experiment. In this paper, simple boreal forest was modeled based on Discrete scatterer model. Through sensitivity study by using the model, we propose several useful indexes to infer vegetation parameters such as soil moisture. This study will be validated by X-band indoor full polarimetric SAR in Niigata University.

Improvement of adaptive-model based decomposition with polarization orientation compensation

Polarimetric SAR decomposition has been used extensively because of its capability to identify the strength of each scattering mechanism from polarimetric SAR data. To achieve higher accuracy, one should take into account not only variation of vegetated terrain but also topography under vegetation. In this paper, we firstly isolate vegetation orientation angle from generalized volume scattering component, and then modify a conventional adaptive model-based decomposition by carefully introducing a concept of topographic polarization orientation compensation (TPOC). Finally, experimental validation is conducted by using P-band AIRSAR image of Black Forest in Germany.

Applicability of SAR to Marine Debris Surveillance After the Great East Japan Earthquake

After the Great East Japan earthquake on March 11, 2011, a huge amount of marine debris of more than 1.5 million tons was generated. It damaged the global marine environment, and reached the west coast of the United States around October 2012. As marine debris spreads with time, the chance of our observation was limited. To achieve an efficient time series analysis, we have to utilize complete advantage of the existing synthetic aperture radar (SAR) data with its wide range of observation conditions. In this study, an applicability of SAR to the tsunami disaster after the Great East Japan earthquake is discussed by estimating several key parameters such as total amount of marine debris and their vector velocities from SAR images obtained just after the earthquake under various observation conditions. Based on the analyses, we also discuss optimum marine debris monitoring by SAR to minimize the damage from tsunami disasters in future.

Effective monitoring for marine debris after the Great East Japan Earthquake by using spaceborne synthetic aperture radar

The Great East Japan Earthquake occurred at March 11, 2011 and caused massive tidal wave. The tsunami swept away a large quantity of rubble and vessels to the sea and they become so-called marine debris. To assess damage situation and protect marine environment, it is essentially required to investigate the status of those marine debris. The technique based on spaceborne synthetic aperture radar (SAR) can be a strong candidate to achieve this ultimate goal, because of its wide observation area and higher resolution with flexible operability: regardless of the day and night and regardless of the weather. We have monitored marine debris on huge amount of spaceborne SAR imagery right after the great disaster and investigated an effective observation of marine debris to predict future direction. In this paper, we firstly define three types of debris as large debris, small debris, and cluster by considering how marine debris looks like on the SAR imagery. Then, an automatic but accurate detection and classification of a large amount of debris on SAR imagery is proposed. Based on those results, resolution and swath width for efficient marine debris monitoring are obtained. Velocity of marine debris is additionally estimated from multi-temporal SAR images to derive optimum swath width.