"The experimental proof of equivalence and rotation angle ambiguity of four-component scattering power decomposition algorithm with rotation of coherency and covariance matrices in SAR polarimetry," J. Remote Sens. Soc. Japan, vol. 33, 2013. (in Press)
The present study shows an experimental proof of the equivalence between the four-component scattering power decomposition (4-CSPD) algorithms based on rotation of covariance matrix and coherency matrix, and the ambiguity in rotation of these matrices. Theoretically, the 4-CSPD algorithms with rotation of the two matrices should be identical. In this paper, an experimental proof is presented for showing the actual equivalence of the two algorithms using polarimetric synthetic aperture radar (POLSAR) data acquired by Phased Array L-band SAR (PALSAR) on board of Advanced Land Observing Satellite (ALOS). An obscure point in the previous publications was also made clear. That is, there is ambiguity in minimizing the cross-polarized term in order to enhance the double-bounce scattering component by rotation of polarimetric matrices. We analyzed how results would be different if this ambiguity remains. Removal of this ambiguity optimally enhances the double-bounce scattering component.
"An analysis of the radar backscatter from oil-covered sea surfaces using moment method and Monte-Carlo simulation: preliminary results," Acta Oceanol. Sin., vol. 32, 2013. (in Press)
An analysis of the radar backscattering from the ocean surface covered by oil spill is presented using a microwave scattering model and Monte-Carlo simulation. In the analysis, a one-dimensional rough sea surface is numerically generated with an ocean waveheight spectrum for a given wind velocity. A two-layered medium is then generated by adding a thin oil layer on the simulated rough sea surface. The electric fields backscattered from the sea surface with two-layered medium are computed with the method of moments (MoM), and the backscattering coefficients are statistically obtained with N independent samples for each oil-spilled surface using the Monte-Carlo technique for various conditions of surface roughness, oil-layer thickness, frequency, polarization and incidence angle. The numerical simulation results are compared with theoretical models for clean sea surfaces and SAR images of an oil-spilled sea surface caused by the Hebei (Hebei province, China) Spirit oil tanker in 2007. Further, conditions for better oil spill extraction are sought by the numerical simulation on the effects of wind speed and oil-layer thickness at different incidence angles on the backscattering coefficients.
"Recent trend and advance of synthetic aperture radar with selected topics," Remote Sens., vol. 5, 2013. (in Press)
The present article is an introductory paper in this special issue on synthetic aperture radar (SAR). A short review is presented on the the recent trend and development of SAR and related techniques with selected topics, including the fields of applications, specifications of airborne and spaceborne SARs, and information contents in and interpretations of amplitude data, interferometric SAR (InSAR) data, and polarimetric SAR (PolSAR) data. The review is by no means extensive, and as such only brief summaries of of each selected topics and key references are provided. For further details, the readers are recommended to read the literature given in the references theirin.
"Extraction of underwater laver cultivation nets by SAR polarimetric entropy," IEEE Geosci. Remote Sens. Lett., vol. 10, no. 2, pp. 231-235, 2013.
This letter describes a technique of extracting and estimating the underwater laver cultivation nets by using the entropy analysis of polarimetric synthetic aperture radar (PolSAR) data. The cultivation nets are placed at 10-20 centimeters below the sea surface, so that the Bragg waves responsible to L-band radar backscatter do not fully develop in this area of effectively shallow water. Consequently, the surface becomes smooth and the backscatter radar cross section (RCS) becomes small in comparison with that from deep water without cultivation nets. If RCS from the cultivation area is at the system noise level, the image can be considered as arising from a random process and the polarimetric entropy should be higher than the open sea area where the radar backscatter is dominated by the single-bounce surface scattering process. We will show that, using the data acquired by Phased-Array L-band SAR (PALSAR) on board Advance Land Observing Satellite (ALOS) over the Tokyo Bay, Japan, the polarimetric entropy is an effective means of extracting underwater cultivation areas in comparison with the amplitude images. The area of the laver cultivation is then estimated by applying a constant false alarm rate (CFAR) to the entropy images to yield good agreement with the ground-truth data.
"Analysis of bar morphology using multi-temporal and multi-sensor satellite images: Example from the Han Estuary, Korea," Marine Geology, 2012 (in Press).
In the present paper, we report a new powerful technique of studying bar morphology using multi-sensor and multi-temporal satellite images with an example from the inaccessible field sites of the Han Estuary in Korea. The technique utilizes the ability of multi-spectral optical sensors for classifying the types of land cover, and the ability of synthetic aperture radar (SAR) for acquiring radar imagery during day and night, irrespective of cloud cover. Combining the data acquired by both types of sensors at different times, the bar morphology of the Han Estuary is analyzed in this study. The Han Estuary is situated on the North Limit Line (NNL) dividing North and South Korea, and considered to have retained its rich ecological value. However, because of the limited access to the estuary, little information is available on the morphology of the Han Estuary, and the technique based on satellite remote sensing described in this paper is the most suitable and only approach to obtaining information on such inaccessible areas. In the present paper, the bar morphology of the Han estuary from the years 2000 to 2009 is investigated using multi-temporal spaceborne optical sensors including those on board of SPOT-5, LANDSAT-7, QuickBird, KOMPSAT-2, and OrbView-3, and also C-band synthetic aperture radar (SAR) on board of RADARSAT-1 and ENVISAT with limited field survey data. Sixteen sets of optical data and twelve sets of SAR data were used to study the spatial and temporal variations of bar morphology and the effects of tide and precipitation. The results show that there was some variation in shape and size of bars depending on the location. The tidal effect is large on the bars closer to the open Yellow Sea, and the effect of precipitation was observed on the bars located in the mouths of rivers. Based on the changes of the bar-water boundary, a topographical map was produced, showing the relative height of bars. Using the technique of satellite remote sensing, this paper has shown, for the first time, the long-term bar morphological characteristics of the restricted Han Estuary on the NLL, using satellite data, and the authors anticipate that the results together with the relative height map of bars will constitute the basis of future morphological and ecological studies in this important estuary.
"Four-component scattering power decomposition algorithm with rotation of covariance matrix using ALOS-PALSAR polarimetric data," Remote Sensing, vol.4, pp.2199-2209, 2012.
The present study introduces the four-component scattering power decomposition (4-CSPD) algorithm with rotation of covariance matrix, and presents an experimental proof of the equivalence between the 4-CSPD algorithms based on rotation of covariance matrix and coherency matrix. From a theoretical point of view, the 4-CSPD algorithms with rotation of the two matrices are identical. Although it seems obvious, no experimental evidence has yet been presented. In this paper, using polarimetric synthetic aperture radar (POLSAR) data acquired by Phased Array L-band SAR (PALSAR) on board of Advanced Land Observing Satellite (ALOS), an experimental proof is presented to show that both algorithms indeed produce identical results.
"Ship recognition by integration of SAR and AIS," J. Navigat., vol.65, vol.2, pp.323 - 337, 2012.
A novel design of an integrated system using Synthetic Aperture Radar (SAR) image and Automatic Identification System (AIS) data is proposed in this paper for the purpose of identifying ships at sea. TerraSAR-X (SpotLight mode) images and AIS data collected over Incheon Port (Korea) and Tokyo Bay (Japan) were used on different dates. Four main steps for integration of SAR and AIS based ships can be identified, namely: 'Time Matching' to retrieve the respective Dead Reckoning (DR) position of the ships at SAR image acquisition times; 'Position Matching' based on nearest neighbourhood re-sampling method with compensation of position shift; 'Size Matching' and 'Speed Matching'. Under each of the matching criteria, the measurement error in each of the matching criteria was found to be less than 20% and the SAR extracted ship's hull boundaries were presented on a screen to display the system results. The results of this study will contribute to the design a Near-Real-Time (NRT) operational system for ship detection, identification, and classification by SARs in different data acquisition modes over various geographical locations at different acquisition times. This novel integrated system design will provide a most important preliminary step towards integration based on ships' hull monitoring in order to recognize 'friend' and 'foe' ship targets over a huge oceanic region and would be useful for coast guards as an early warning system.
"Assessment of building damage in 2008 Wenchuan earthquake from multi-temporal SAR images using Getis statistic," IEICE Trans. Commun., vol.E94-B, no.11, pp.2983-2986, 2011.
In this paper, the Getis statistic is applied to ALOS- PALSAR (Advanced Land Ovserving Satellite-Phased Array L-band Synthetic Aperture Radar) images for assessing the building damage caused by the Wenchuan earthquake in 2008. As a proposed image analysis, a simulated building image using mapping and projection algorithm is first presented for analysis of the Getis statistic. The results show the high accuracy of the assessment of the proposed approach. The Getis statistic is then applied to two ALOS-PALSAR images acquired before and after the Wenchuan earthquake to assess the level of building damage. Results of the Getis statistic show that the damage level is approximately 81%.
"Polarimetric scattering of sea ice and snow using L-band quad-polarized Palsar data in Kongsfjorden, Svalbard (in Korean)," Ocean and Polar Research, vol.33, no.1, pp.1-11, March, 2011.
This study describes measurements of fast ice recorded on May 23, 2009, in Kongsfjorden (translated as 'Kongs Fjord'), an inlet on the west coast of Spitsbergen in the Svalbard Archipelago. Seasonal fast ice is an important feature for Svalbard fjords, both in relation to their physical environment and also the local ecosystem, since it grows seaward from the coast and remains in place throughout the winter. Ice thickness, snow, ice properties, and wind speed were measured, while SAR (Synthetic Aperture Radar) data was simultaneously observed two times from ALOS-PALSAR (L-band). Ice thickness was 25-35cm while the thickness of ice floe broken from fast ice was 10-15cm. Average salinity was 1.9-2.0 ppt during the melting period. Polarimetric data was used to extract H/A/alpha-angle parameters of fast ice, ice floe, snow and glacier, which was classified into 18 classes based on these parameters. It was established that the area of fast ice represents surface scattering which indicates low and medium entropy surface scatters such as Bragg and random surfaces, while fast ice covered with snow belongs to a zone of low entropy surface scattering similar to snow-covered land surfaces. The results of this study will contribute to various interpretations of interrelationships between H/A/alpha parameters and the scattering processes of sea ice.
"A simple moment method of forest biomass estimation from non-Gaussian texture information by high-resolution polarimetric SAR," IEEE Geosci. Remote Sens. Lett., vol.7, no.4, pp.811-815, October, 2010.
A simple and robust method is described to estimate forest biomass by high-resolution polarimetric synthetic aperture radar (SAR). The method is based on the regression analysis between the measured biomass from the ground survey and the second intensity moment of the non-Gaussian texture in the cross-polarized L-band SAR images. The SAR data used in the analysis were acquired by the airborne Pi-SAR over the coniferous forest in Hokkaido, Japan. The regression analysis was first carried out, and a model function was derived to relate the intensity moment and the measured biomass in 19 forest stands. Using this model function, the biomass values were estimated and compared with those of 21 different stands with known biomass. The average accuracy of the moment model was found to be 85% which is similar to that of the previous K-distribution model. The advantage of this method over the distribution-based model is that there is no need to search a specific distribution function which fits best to the image texture.
"On a novel approach using MLCC and CFAR for the improvement of ship detection by synthetic aperture radar," IEEE Geosci. Remote Sens. Lett., vol.7, no.2, pp.391-395, April, 2010.
MLCC (Multi-Look Cross-Correlation) is a useful technique to extract the images of ships embedded in heavy sea clutter by SAR (Synthetic Aperture Radar). In the ship detection experiment in 2006 by PALSAR (Phased Array L-band SAR) on board of ALOS (Advance Land Observing Satellite), we applied MLCC to PALSAR data in order to extract small fishing boats. The result was that some boats were detected by thresholding MLCC coherence images under favorable conditions. However, it was also found that the threshold method was not suitable to automatically determine the threshold levels corresponding to desired FAR (False Alarm Rate) values. In order to overcome this problem and to improve the accuracy of ship detection by MLCC, we propose a new and simple technique of MLCC-CFAR (or Gamma-CFAR). In this method, CFAR (Constant False Alarm Rate) is applied to inter-look coherence image produced by MLCC. We tested this method using simulation and PALSAR data, and found substantial improvement in SNR (Signal to Noise Ratio) and FAR in comparison with the coherence image alone. In this letter, we summarize the MLCC-CFAR algorithm and the experimental results.
"Estimating spatial distribution of oceanic waves using polarimetric synthetic aperture radar (in Japanese)," JSCE J. Hydraulic Coastal Environ. Engineer., vol.66, no.1, pp.1416-1420, October, 2010.
A new technique for estimating spatial distribution of oceanic waves using polarization ratio data of PALSAR synthetic aperture radar is presented. Polarization ratio of the backscatterd power from the ocean surface at vertical to that at horizontal polarization (VV/HH) is closely correlated to the local incidence angle which reflects local water surface slope. At first linear relationship between incidence angle Ζand polarization ratio Pwas derived from the trend of the entire target aria. Then the pixels on the polarization ratio image correspond to the steepest slope points were selected with reference to the HH backscatterd power image. The mean steepest slope of ocean surface was estimated from the selected pixels and the linear relation between Ζand P. Finally, assuming sinusoidal waves, mean wave height was estimated from the steepest slope and mean wave length estimated from the HH image. This method was applied to each divided 3x3km subareas and spatial distribution of wave height was estimated.
"Comparison and evaluation of ship detection and identification algorithms using small boats and ALOS-PALSAR," IEICE Trans. Commun., vol.E92-B, no.12, pp.3883-3892, December 1, 2009.
The final goal of the present project is to develop a ship detection and identification system by integrating spaceborne synthetic aperture radar (SAR), ground-based maritime radar and automatic identification system (AIS); and this article presents the results of the first phase experiments and current status toward achieving this goal. The data acquired by the Phased Array L-band SAR (PALSAR) on board of the Advanced Land Observing Satellite (ALOS) were used as SAR data, and X-band maritime radar including AIS were used as a ground-based system. The work is divided into two experimental phases. The first phase is to examine the ability of PALSAR to detect ships whose sizes are comparable with the SAR resolution cells, and the second is to incorporate the PALSAR data with those acquired by the ground-based radar with AIS. For the experiments in the first phase, we deployed three small fishing boats whose lengths ranged from approximately 8 m to 15 m in the Tosa Bay in Kochi, Japan in 2006. The experiments were carried out for four observation PALSAR modes, including FBS (Fine Beam Single) 34.3, FBS 21.5, FBD (Fine Beam Double) 41.5, and PLR (PoLaRimetric) 20.5, where the numbers in each modes represent the off-nadir angles. For extracting the boats from the PALSAR images, five algorithms were considered. These algorithms were the amplitude-based, CFAR (Constant False Alarm Rate), MLCC (Multi-Look Cross-Correlation), CCF (Cross-Correlation Function) of HH- and HV-polarization amplitudes, and polarimetric analyses. This paper summarizes the results of the first phase experiments; the summary of the integrated system in the second phase will be reported in near future.
"Estimation of bridge height over water from polarimetric SAR image data using mapping and projection algorithm and de-orientation theory," IEICE Trans. Commun., vol.E92-B, no.12, pp.3875-3882, December 1, 2009.
An inversion method of bridge height over water by polarimetric synthetic aperture radar (SAR) is developed. A geometric ray description is illustrated for the scattering mechanisms from a bridge over water surface. Based on Mapping and Projecting Algorithm (MPA), polarimetric SAR images of a bridge are simulated, and the results show that the bridge in the SAR images can be identified by three strips corresponding to single-, double-, and triple-scattering processes respectively. A set of parameters based on the de-orientation theory is applied to locate the image positions of the three scattering processes from the bridge structure. The thinning and clustering algorithms and Hough transform are employed to detect the lines indicating different types of scattering. Then, these lines are used to invert the bridge height. Fully polarimetric data of airborne Pi-SAR at X-band are applied to the inversion of the height and width of the Naruto Bridge in Japan, and also the spaceborne ALOS-PALSAR single-polarization data are used to estimate these parameters of the Eastern Ocean Bridge in China. The results show a good feasibility of the bridge height inversion by the proposed method.
"Accuracy of the K-distribution regression model for forest biomass estimation by high-resolution polarimetric SAR: Comparison of model estimation and field data," IEEE Trans. Geosci. Remote Sens., vol.46, no.4, pp.1058-1064, April, 2008.
In our previous study, it was shown that non-Gaussian amplitude fluctuations in high-resolution polarimetric SAR (Synthetic Aperture Radar) data of coniferous forests can accurately be described by the K-distribution, and that the order parameter of the K-distribution can be a new useful parameter for estimating the tree biomass of coniferous forests from L-band cross-polarization amplitude images in a wider range than the conventional method using the RCS (radar cross section) alone. The result was based on the analysis of the ``ground-truth'' biomass data of 19 forest stands and the airborne Pi-SAR (Polarimetric interferometric SAR) L-band data over the Tomakomai forests in Hokkaido, Japan. From this relation, an empirical regression model was developed for estimating forest biomass from SAR data. In the present paper, we report the results on the further analyses of this regression model. The validity of the K-distribution is first reconfirmed using the AIC (Akaike Information Criterion), followed by the description on the accuracy of the model. To examine the model accuracy, we carried out further field measurements on 22 forest stands in 2005, and the ground survey was made in 2006 to find out the causes of several anomalous data. From the comparison of the model-based biomass and ground-truth data, it is found that the accuracy of the regression model is approximately 86%. The regression model is then updated for practical applications to estimating the Hokkaido forests by including the ground-truth data of all 41 forest stands, and limitations of the model are discussed.
"Study on ship detection using SAR dual-polarization data: ENVISAT ASAR AP mode," Korean J. Remote Sens., vol.24, no.5, pp.445-452, October, 2008.
Preliminary results are reported on ship detection using coherence images computed from cross-correlating images of multi-look-processed dual-polarization data (HH and HV) of ENVISAT ASAR. The traditional techniques of ship detection by radars such as CFAR (Constant False Alarm Rate) rely on the amplitude data, and therefore the detection tends to become difficult when the amplitudes of ships images are at similar level as the mean amplitude of surrounding sea clutter. The proposed method utilizes the property that the multi-look images of ships are correlated with each other. Because the inter-look images of sea surface are covered by uncorrelated speckle, cross-correlation of multi-look images yields the different degrees of coherence between the images and water. In this paper, the polarimetric information of ships, land and intertidal zone are first compared based on the cross-correlation between HH and HV images. In the next step, we examine the technique when the dual-polarization data are split into two multi-look images. It was shown that the inter-look cross-correlation method could be applicable in the performance improvement of small ship detection and the land masking. It was also found that a simple combination of coherence images from each co-polarised (HH) inter-look and cross-polarised (HV) inter-look data can provide much higher target-detection possibilities.
"Estimation of forest biomass and experimental validation based on Pi-SAR polarimetric data and K-distribution index," (in Chinese) J. Remote Sens., vol.12, no.3, pp.477-482, 2008.
Employing Pi-SAR polarimetric data acquired in 2002 and 2003, forest biomass estimation approach is studied on Tomakomai forests located in Hokkaido, Japan. The purpose of this project is to develop effective approach for estimating forest biomass. The ground truth data of 19 test sites are in hand. In this test sites, one sample stand of 20m~20m are selected and tree height, age, basal area, diameter of breast height and tree species are measured, the biomass is then calculated. The conventional Radar cross section (RCS) method is first investigated. It is found that RCS increases with biomass and becomes saturated rapidly, for the data used in this paper, the L-band RCS saturation levels are found approximately to the biomass of 40 tons/ha, the tree age of 30 years, the basal area of 30 m2/ha, and the tree height of 8m. The RCS saturates at 20 tons/ha for X-band data. Therefore, forest biomass beyond saturation level cannot be estimated utilizing RCS. To search the quantitative relation between high-resolution SAR data and forest parameters, statistical analysis approach is utilized. The probability density function of image amplitude is then investigated, and among Rayleigh, log-normal, Weibull and K-distributions, the K-distribution is found to fit best to the L-band data of all polarizations according to the Akaike information criterion (AIC). The relations between K-distribution index and tree parameters including biomass, tree age, height, basal area, are investigated. It is found that the tree biomass correlated best with the index parameter. Moreover, K-distribution index increases with biomass beyond RCS saturation level, and the highest correlation coefficient is obtained at cross-polarization. The regression model is developed between K-distribution index and forest biomass at cross-polarization based on 19 test sites data. In 2005 August and September, we further collected ground truth data of 23 test sites. Utilizing the relation of K-distribution index of cross-polarization and forest biomass, the biomass estimation is made for the 23 test sites. The comparison of estimated biomass and measured ground truth data is made, and the results show that the average accuracy of the estimation reaches 85%. It is concluded that, at least for the Hokkaido forests, this empirical model is can effective and superior way of estimating forest biomass from polarimetric SAR data in comparison with the conventional RCS model.
"In Search of the Statistical Properties of High-Resolution Polarimetric SAR Data for the Measurements of Forest Biomass beyond the RCS Saturation Limits," IEEE Trans. Geosci. Remote Sens. Lett., vol.3, no.4, pp.495-499, October, 2006.
The purpose of this letter is to present the results on the study of searching effective parameters that describe the relation between high-resolution SAR (synthetic aperture radar) images and forest parameters. The study is based on the non-Gaussian texture analysis of the polarimetric airborne Pi-SAR data over coniferous forests in Hokkaido, Japan. The RCS (radar cross section) in terms of forest biomass is first analysed. It is found that the L-band RCS increases steadily with biomass and saturates at approximately 40 tons/ha. These results are similar to the previous studies. The probability density function (PDF) of image amplitude is then investigated, and among Rayleigh, log-normal, Weibull and K-distributions, the K-distribution is found to fit best to the L-band data of all polarizations, although the Weibull distribution fits equally well. Further, the correlation between the tree biomass and the order parameter of the K-distribution in the cross-polarization images is found to be very high, and the order parameter increases consistently with biomass to approximately 100 tons/ha which is well beyond the saturation limit of the L-band RCS. Thus, the order parameter of the K-distribution can be a promising new parameter to estimate forest biomass from high-resolution polarimetric SAR data in a much wider range than the conventional RCS method.
"On the Bragg scattering observed in L-band synthetic aperture radar images of flooded rice fields," IEICE Trans. Commun., vol.E89-B, no.8, pp. 2218-2225, August 1, 2006.
This article presents the analysis of the Bragg scattering phenomenon which has been observed in the images of machine-planted rice paddies acquired by the JERS-1 L-band synthetic aperture radar (SAR). The simultaneous measurements of rice plants were made at the SAR data acquisition times. Large differences of 20-25 dB in image intensity between the transplanting and ripening stages are found to be dependent on the planting direction and bunch separation. This selective image enhancement is a result of the Braggresonance backscatter due to the double-bounce of incident L-band microwave between the flooded water surface and periodically plantedbunches of rice plants. Support for the idea of double-bounce scattering is provided by the decomposition analysis of L-band and X-band polarimetric Pi-SAR data; and a simple numerical simulation basedon the physical optics model shows fairly good agreement with the JERS-1 SAR data. The results presented in this paper is mainly of academic interest, but a suggestion can be made on the selection of suitable microwave band for monitoring rice fields.
"Recent Trend and Information Extraction from Synthetic Aperture Radar Images," (in Japanese) IEICE Trans. Commun., vol.J89-B, no.7, pp. 1024-1035, July 1, 2006.
In recent years, there has been a rapid development in SAR (Synthetic Aperture Radar) technology, extending from theconventional single-wavelength single-polarization image forming radar to InSAR (Interferometric SAR) which measures surface height variation and crust movements on land and current velocity over ocean, and PolSAR (Polarimetric SAR) that provides image classification techniques with improved high accuracy. At present, SAR remote sensing is a state-of-the-art technology making full use of highly advanced information processing, and it is an essential measurement technique in the field of geoscience. In the present article, accounts are given on the techniques of information extraction from SAR amplitude images, InSAR and PolSAR data, and on the applications to earth science as well as the recent research trend.
"Single look classification accuracy for polarimetric SAR," Int. J. Remote Sens., vol.27, no.22, pp. 5073-5080, November, 2006.
Polarimetric SAR systems such as ALOS-PALSAR and Radarsat-2 can operate in many different modes. The use of additional polarisations may require additional time and operating power and it is important to justify this by an improvement in final classification accuracy. A fully polarimetric, dual frequency AirSAR scene from a rice growing area in Japan is classified by a maximum likelihood method based on the Wishart distribution. It is shown how the measured covariance matrices determine the separation accuracy between two classes. Closed form expressions are then given for the expected single look accuracy of the maximum likelihood classifier as a function of the class covariance matrices. This can be used to quickly compare the high resolution classification performance for different polarimetric systems to decide on a particular operating mode.
"Interlook cross-correlation function of speckle in SAR images of sea surface processed with partially overlapped subapertures," IEEE Trans, Geosci. Remote Sens., vol.43, no.4, pp.695-701, April, 2005 .
In the present paper, a general integral expression is derived and discussed for the cross-correlation function (CCF) of speckle patterns in synthetic aperture radar (SAR) images processed by using partially overlapped subapertures of arbitrary Doppler center frequencies (or equivalent azimuth times). It is shown that, under the white noise approximation for the backscattered field, the CCF of the interlook speckle intensity patterns is given by the squared modulus of the autocorrelation function of the amplitude weighting function of subapertures where the time lag is the center time difference. It is also shown that the CCF of the interlook speckle patterns is independent of the surface coherence time of sea surface. The integral expression for the intensity CCF is then evaluated for a rectangular weighting function, and comparison is made with the JERS-1 L-band and RADARSAT-1 C-band SAR images of sea surface to test the theory. The CCFs computed from the JERS-1 SAR data show excellent agreement with the theory, and good agreement is obtained with the RADARSAT-1 data.
"Ship Detection Based on Coherence Image Derived from Cross Correlation of Multilook SAR Images," IEEE Geosci. Remote Sens. Lett., vol.1, no.3, pp.184-187, July, 2004.
A new simple technique is presented to extract ships from synthetic aperture radar (SAR) images. The procedure is to compute the cross-correlation values between two images extracted by moving windows of a small size from the multilook SAR intensity (or amplitude) images. A coherence image, consisting of the cross-correlation values of the intensity images, is then produced. Ships are deterministic targets, so that their interlook subimages possess higher degree of coherence than the uncorrelated random images of the surrounding sea surface. The main advantage of this method over the conventional CFAR (constant false alarm rate) is its ability to detect, under favorable conditions, ``invisible'' images of ships embedded in the speckled image of the sea surface. The technique is tested using a RADARSAT-1 image in which a known and several unknown ships are present. The use of complex images and the exploitation of short decorrelation times of small-scale ocean waves to obtain an extra look are also discussed.
"Rice status and microwave characteristics - Analysis of rice paddy fields at Kojima Bay using multi-frequency and polarimetric Pi-SAR radar data images -," (in Japanese) J. Remote Sens.Soc. Japan, vol.23, no.5, pp.473-490, December, 2003.
South-east Asia has a rainy-season at the crop growing period, and it is difficult to observe agricultural land in this season using optical remote sensing. Synthetic Aperture Radar (SAR) can observe the earth's surface without being influenced by clouds. However, it is less useful for observing agricultural land, because satellite SAR has only one data band. Recently, SAR is able to provide multi band and multi polarimetric dara. Pi-SAR, an airborne SAR developed by NASDA and CRL, can provide L and X bands and fully polarimetric data. Rice is the main crop in Asia, and we studied the chracteristicmicrowave scatter on rice paddy fields using Pi-SAR data. Our studyarea was the rice paddy fields in Kojima reclaimed land in Japan. We had two fully polarimetric data sets from 13 July 1999 and 4 October 2000. First, we processed the color polarimetric composite image. Next we calibrated the phase of each polarimetric data using river area by the Kimura method. After that we performed decompositionanalysis and drew polarimetric signatures for understanding the status of rice paddy fields. At the rice planting period, rice paddy fields are filled with water and rice plants are very small. The SARmicrowave scatters on water surfaces like a mirror, called 'mirror (or specular) reflection'. This phenomenon makes backscatter a small value at the water-covered area. The image from July is about one month after trans-planting and rice plants are 20-40 cm in height. X-band microwave scatters on the rice surface, but L-band microwave passes through rice bodies and shows mirror reflection on water surfaces. Some strong backscatter occurs on rice paddy fields especially VV polarization because of bragg scattering. The fields where bragg scattering returns strong VV scatter because the space between rice stems cause resonation in the L-band wavelength. We can easily understand bragg scatter by using polarimetric data. Using theimage from October at just before harvest, L-band polarimetric datacan detect various rice statuses such as standing, inclining, or lying. We conclude that multi band and fully polarimetric SAR data can quantity detect crop growth, as do optical sensors in all weatherconditions.
"Segmentation of SAR images using multitemporal information,'' IEE Proc. Radar, Sonar and Navigation, vol.150, no.5, pp.367-374, October, 2003.
The maximum likelihood method of SAR segmentation has the potential to retain single pixel accuracy without requiring heuristic decisions. Normally a probabilistic measure is used to merge individual regions without assuming any prior knowledge for the underlying cross-sections. However, for a reasonable multitemporal scene there may beconsiderable information available from the varying cross-sections over time. An example is given where this information can be extracted by an initial classification. It is then shown how the segmentation scheme can be modified to incorporate this information via anestimate of the multitemporal underlying class distributions. Usingsingle-look Radarsat data at 8 m resolution, it is demonstrated howthe final segment population can be significantly reduced. From a comparison with ground survey data and a high-resolution AirSAR image, the structural quality of the segmentation is shown to be improved.
"Non-uniform azimuth image shift observed in the Radarsat images of ships in motion," IEEE Trans. Geosci. Remote Sens., vol.40, no.10, pp.2188-2193, October, 2002.
This paper describes, for the first time to the authors' knowledge, the phenomenon of non-uniform azimuth image shift of a rigid body observed in the Radarsat SAR (synthetic aperture radar) images of cruising ships. The effect is caused by the different slant-range velocities of coherent scatterers across the hull associated with the ship motions. The slant-range velocity is estimated from the SAR image of an identified ship, and compared with that computed from a numerical model using the ship's specification and meteorological data. The result indicates that the dominant contribution to the non-uniform image shift is the pitching motion of the ship. Further comparison is made with the wave orbital velocity, and the results are shown to be in good agreement. Comparisons are also made between the SAR-derived slant-range velocities of two unknown ships and wave orbital velocities, and reasonable agreement is obtained. One of the ships' images exhibits not only non-uniform shift but also image skew. The latter skewing effect may be caused by rolling of the ship.
"A theory on the distribution function of backscatter radar cross section from ocean waves of individual wavelength," IEEE Trans. Geosci. Remote Sens., vol.38, no.2, pp.811-822, March, 2000.
A new and simple method is presented of interpreting the `distribution' of the backscatter radar cross section (RCS) from ocean waves of individual wavelength. Using the Kirchhoff scattering (Physical Optics) theory, the `cumulative' RCS from the ambient waveheight spectrum is first computed as a function of the wavenumber. Differentiating this cumulative RCS yields the distribution function of the RCS from ocean waves of different wavenumbers. The present study shows, as a general feature, that the maximum RCS contribution tends to come from the ocean waves of wavenumbers closer to the Bragg wavenumber as the radar incidence angle increases. Under low wind speeds (` 2m/s) and at the incidence angles between 20 and 45, the dominant RCS contribution at both C- and L-bands comes from ocean waves of wavenumbers close to the Bragg wavenumber. Under intermediate to high wind speeds (`10-15 m/s) and at the small incidence angle of 20, the radar backscatter is dominated by long ocean waves with little contribution from the Bragg waves at both bands. At the intermediate incidence angles (`35to 45), the RCS distribution is centered at wavenumbers near the Bragg wavenumber at all wind speeds from 2 m/s to 15 m/s. These features are more pronounced at C-band than L-band. The present theory is based on the Kirchhoff scattering model, and as such, its validity may also be limited to a range of small to intermediate incidence angles where the polarization effect is not very significant. The RCS distribution is evaluated for the ambient sea surface and the surface perturbed by varying currents induced by the interaction between the current and bottom topography, and the consequence of the result is discussed.
"Determination of ocean wave propagation direction by split-look processing using JERS-1 SAR data," IEEE Trans. Geosci. Remote Sens., vol.37, no.2, pp.849-855, March, 1999.
The purpose of the present article is to show that the JERS-1 synthetic aperture radar (SAR), primarily designed for global land observation, can be applied to deriving ocean wave spectra and determining the wave propagation direction. Split-look (multilook) processing is used to produce a sequence of individual sub-images of ocean waves. This processing technique has an inherent property that the successive sub-images are formed at different discretely delayed times. Wave propagation direction can, therefore, be determined from a pair of wave images in different looks. The L-band JERS-1 SAR, despite its low signal-to-noise ratio, has an advantage over C-band spaceborne SARs for having a longer integration time, suitable for the wave analysis using split-look processing. Several different methods are applied to the wave data collected near the coasts of Japan, including the methods of look sum (taking the standard Fourier spectrum of look-summed image), spectral sum, spectral phase shift, and cross-correlation function of sub-images.
"Multi-Frequency SAR images of ship-generated internal waves," Int. J. Remote Sens., vol.18, no.18, pp.3709-3718, December, 1997.
During the joint U.K./U.S.A. experiment conducted in the Loch Linnhe, Scotland, U.K., in 1989, the Jet Propulsion Laboratory (JPL) multi-frequency airborne synthetic aperture radar (SAR) was deployed to investigate the relation between the SAR images and ship-generated internal waves. One of the intriguing results, consistent throughout the experiment, was the striking difference in phase (position) between the images of internal wave wakes in the P-band (`68 cm) wavelength and those in L-band (`24 cm) and C-band (`6 cm) wavelengths. An explanation for this difference is found in the sensitivity of different radar wavelengths to oceanic surface waves in different ranges of wavelengths, that are perturbed by varying surface currents by different amounts and at different positions, depending on the wavelength of surface waves.
"Synthetic aperture radar and its applications I : principles of synthetic aperture radar," (in Japanese), J. Visualization Soc. Japan, vol.15, no.56, pp.41-48, January, 1995.
"Synthetic aperture radar and its applications II : applications to remote sensing of oceans," (in Japanese), J. Visualization Soc. Japan, vol.15, no.57, pp.118-125, April, 1995.
@SAR (Synthetic Aperture Radar) is an imaging radar capable of forming the two-dimensional high-resolution images of land and sea at microwave bands of wavelengths ranging from a few mm to a meter. It is one of the most important sensors for the present and future remote sensing. SAR is also called "microwave holography" at its developing stage between 1960's to 1970's , and its principle is to generate a microwave, instead of optical, hologram and to reconstruct an image by means of optical processing. The research at that period was of a small-scale utilizing airborne platforms. The L-band SAR on board of the marine observation satellite "SEASAT' launched by NASA in 1978 provided fine images of land and sea and the vast amount of potential information and variety of possible applications. Since then, SAR has attracted much attention, and large-scale experiments and research have been carried out, using the shuttle and airborne platforms. At present, SARs on board of the ERS-1 (ESA), JERS-1 (Japan) and ALMAZ-1 (Russia) are collecting data, and further projects toward the 21st century involving SARs are being planned by several organizations. Furthermore, due to the rapid advance of computer technology, almost all SAR images are produced using digital processors. An advantage of digital processing is to obtain the image in a complex format, enabling us to directly access the phase information that cannot be obtained from optically processed images. The phase is said to contain information much more than the amplitude information. In fact, SAR interferometry, which is a focus of current attention, is a technique utilizing this phase information. There are three reasons why SAR is such an important sensor. The first is to achieve high-resolution in the azimuth (along-track) direction by means of aperture synthesis, which cannot be attained by real aperture radars. The second reason is that SAR operating at the microwave band has a day-and-night and all-weather imaging capability, irrespective of the Sun's illumination and cloud cover. This ability is particularly useful in the Polar Regions where the half a year is in continuous darkness, and in the tropical rainforests under continuous cloud cover. The third reason is that the information at the microwave band is different from that at the optical and infra-red bands. Although SAR alone can have its advantage to provide much information, the accuracy of extracting and evaluating potential information will increase considerably by combining, integrating and analyzing the SAR data with those acquired from other sensors having different parameters.
@Remote sensing of oceans by SAR is one of the important research fields for the present and future Earth environment. The sea occupies two third of the Earth's surface, affecting climate, and provides a huge amount of resources for the human race. However, because of its vast nature, the amount of data collected by ships and buoys is partial and the applications are also limited. On the other hand, satellites can collect information on a global scale, and by taking into account the characteristics of microwave, spaceborne SARs are essential for remote sensing of oceans.
@In the first part of the present tutorial review, the principle of SAR is illustrated. The pulse compression technique to achieve fine range (cross-track) resolution is similar to the technique of aperture synthesis, and since the theory is not very complicated, some mathematical descriptions are provided. However, those who do not require such details can skip to the summary section where the basic method is explained without using the technical mathematics. Ocean surface is dynamic and the SAR image is affected inherently by the motion of scatterers; hence the effects of the scatterer's motion is taken into consideration.
@In the second part, SAR applications to ocean remote sensing are illustrated by describing oceanic phenomena using real images. These features include surface and internal waves, bottom topography, ocean currents, fronts, large-scale eddies, wind, cold and warm waters, sea ice and icebergs. It is not possible in this limited space to describe the detailed process of SAR imagery and to show many images, and therefore interested readers can refer to some key and important references provided in this paper.
"Resolution of a controversy surrounding the focusing mechanisms of synthetic aperture radar images of ocean waves," IEEE Trans. Geosci. Remote Sens., vol. GE-32, no.5, pp.1004-1016, September, 1994.
This paper addresses key problems regarding the focusing of synthetic aperture radar (SAR) images of ocean surface waves, explaining why applying a processor defocus will generally yield an enhanced image, why the same defocus applies to both image modulations brought about by the radar cross section and by the velocity bunching process, and why the effects apply to both single-look and multilook systems independently of look relocation. Two interpretations are given for the case when surface scatterers are stationary, but modulated in reflectivity (radar cross section) by a propagating wavefield. The first interpretation is what will be called a "degrade-and-shift" model. In it, a processor focusing adjustment degrades a point image. However, the overall image can be enhanced because an appropriate defocus results in a shifting of points in such a way that the image can most closely resemble the image of the time-invariant (or "frozen") reflectivity. The second interpretation is a "defocus-and-refocus" model in which the image of a time-varying reflectivity is defocused and may be refocused to enhance the image. In justifying this "defocus-and-refocus" model, it is shown that the radar return from stationary scatterers of time-varying reflectivities is identical to that from physically moving scatterers of constant reflectivity. Thus, the two interpretations are not contradictory; they are, fundamentally, equivalent. The models support the use of a processor defocus corresponding to one half the wave phase velocity. Both qualitative and quantitative illustrations of the effects are given. Finally, it is shown that the same defocusing effect applies to image modulations brought about by the velocity bunching process.
"Modulation of waveheight spectrum and radar cross section by varying surface currents," IEEE Trans. Geosci. Remote Sens., vol. GE-32, no.5, pp.995-1003, September, 1994.
This paper investigates the validity of the relaxation model which is generally used to describe the modulation of wind-generated ocean waves by slowly varying surface currents, and the applicability of the model to illustrating features observed in the radar images of internal wave wakes. The amplitude and phase modulations in the waveheight spectrum are examined for sinusoidal current variations: these changes are then translated into the radar cross section (RCS) modulation using the Kirchhoff scattering theory. The present study concludes that: 1) the relaxation model is valid under intermediate to strong wind conditions and/or weak current signatures, and the conditions for the validity are given; 2) the relaxation model combined with the Kirchhoff model could, under certain conditions, explain the differences in amplitude and phase between the P -band and L (C - and X -) band SAR images of ship-generated internal wave wakes observed during the 1989 Loch Linnhe experiment; and 3) the directional dependence of the relaxation rate enhances these differences.
""Microwave remote sensing of the ocean: a review," J. Remote Sens. Soc. Japan, vol.12, no.1, pp.92-101, March, 1992.
This paper is a short review on the principles and applications of the active microwave remote sensing of the ocean, focusing on the synthetic aperture radar observation of sea surface.
"Dependence of SAR azimuth image displacement of range moving scatterers on processor focal setting," IEEE Trans. Geosci. Remote Sens., vol. GE-30, no.1, pp.180-184, January,1992.
This paper is a reappraisal of the principal theory on the azimuth displacement of the synthetic aperture radar (SAR) image of a point scatterer moving in the range direction with a constant velocity. New results are presented which describe the dependence of the image shift on the processor focal setting. It is shown that, provided the bandwidth of the processor is greater than or comparable with the bandwidth of the signal, there is a strong dependence of the azimuth image shift on the processor focusing. Simple geometric illustrations based on the optical processor are provided to interpret the effects.
"Statistical analysis of azimuth streaks observed in digitally processed CASSIE imagery of sea surface," IEEE Trans. Geosci. Remote Sens., vol. GE-29, no.5, pp.727-735, September, 1991.
During the Canberra SAR Sea Imaging Experiment in 1987 (CASSIE'87), numerous azimuth streaks were observed by the X-band synthetic aperture radar (SAR). The data are digitally processed, enabling us to make accurate statistical analysis. The probability density function (PDF) and higher order intensity moments are first calculated and compared with theory for both the image areas containing a few and a large number of streaks. The results show that the PDF corresponding to the area of a few streaks fits well with a log-normal distribution, while the image intensity containing many streaks is found to obey K-distribution. The two-dimensional autocorrelation function (ACF) is shown to consist of a central peak followed by a slow decay. The presence of a slow decay in range direction indicates that the backscattered radar cross section (RCS) is spatially correlated over the extent of breaking waves that are considered to be the source of azimuth streaks. The slow decay in azimuth direction is more pronounced than that in range direction due to the scatterers' motions in addition to the spatial correlation of RCS. The comparison of the ACF's of the streaked images processed at different processor focal settings suggests that the dominant motion effect is caused by the scatterers' random motion, although a little dependence, but not dominant, has been observed to indicate the presence of uniform scatterers' motions and/or unknown aircraft motion errors. Based on the experimental observations, a theoretical model is postulated to account for the correlated non-Rayleigh scattering amplitude statistics and the random motion of scatterers. Attempts are then made to recover the range ACF of RCS from the image ACF and to predict the azimuth component of the image ACF.
"Speckle cross-correlation function in multilook SAR images of moving discrete scatterers," Int. J. Remote Sens., vol.12, no.9, pp.1933-1946, December, 1991.
The properties are described of the cross-correlation function (CCF) of the non-Gaussian speckle in the images of a moving random rough surface produced by a synthetic aperture radar (SAR) operating in the multilook mode. The scattering surface is considered to contain discrete scatterers, giving rise to the backscattered complex field which consists of spiky non-Gaussian random amplitude fluctuations. This spatial complex field is subjected to temporal changes due to the scatterers. motions which include random motion, uniform azimuth velocity and uniform range acceleration, since they are the major components to cause the motion effects which appear in the CCF. This type of scattering model can be found in the radio wave backscatter from breaking or near-breaking sea waves. The multilook impulse responses from a moving point scatterer are briefly described; they are then convolved with the backscattered spatial field to yield the subimages of different looks. A general integral form for the correlation function of image speckle intensity is expressed in terms of the degraded multilook impulse responses and the number of discrete scattering amplitude fluctuations. Explicit expressions of the speckle CCF between the subimages are analysed for different scatterers' motions, highlighting the motion dependent effects inherent to multilook SARs.
"Autocorrelation function of non-Gaussian speckle in synthetic aperture radar images of dynamic sea surfaces," Int. J. Remote Sens., vol.12, no.2, pp.315-327, February,1991.
This paper presents a detailed theory on the statistical properties of non-Gaussian speckle in the synthetic aperture radar (SAR) images of dynamic sea surfaces, where breaking or near-breaking waves are present. The backscattered complex field is modelled as the sum of two independent fields: one arises from breaking waves which occur discretely and therefore the amplitude fluctuations are also discrete, and the other represents the weak background field of an uniform amplitude due to non-breaking sea waves and/or system noise. The elementary scatterers are considered to be either in random motions, and the magnitude of the motion is considered to be much greater for breaking waves than the background non-breaking waves. This compound backscattered field is then imaged by a SAR, resulting in a non-Gaussian speckle pattern which appears as the pronounced azimuth streaks. A general and simple expression is derived for the autocorrelation function (ACF) of the speckle intensity in terms of the statistical properties of the spatial backscattered field and the impulse response degraded by the scatterers' motions. The ACF is evaluated and discussed for a Gaussian impulse response and for different scatterers' motions.
"Synthetic aperture radar imagery of range travelling ocean waves," IEEE Trans. Geosci. Remote Sens., vol. GE-26, no.1, pp. 30-37, January, 1988.
In synthetic aperture radar (SAR) imagery of ocean waves, an image modulation by radar foreshortening exists in addition to backscatter radar cross sections. The foreshortening effect is due to surface height differences making sections of a rough surface fall into different range bins than they would if the surface were flat. Since the degree of this modulation changes according to the local waveheight, foreshortening alone could produce the image pattern corresponding to the wave field. The effect termed as "range bunching" is maximum for range traveling ocean waves and vanishes for azimuth traveling waves. It also decreases with decreasing radar look angle. In this paper the imaging processes of range traveling waves are investigated by taking account of both the cross section and range bunching. The SAR transfer function and image modulation functions are defined to describe the relative importance and the coupling effects of the two contributions. It is shown that in low to moderate sea states where the major variation in backscatter arises from local surface tilt, the image modulation by the cross section is enhanced by range bunching and the effect increases with increasing wave slope and also with decreasing look angle. For ocean waves in high sea states with a small variation in the cross section and steep wave slopes, range bunching may become an important mechanism for the interpretation of the images formed by SAR with small look angles.
"Dependence of speckle statistics on backscatter cross-section fluctuations in synthetic aperture radar images of rough surfaces," IEEE Trans. Geosci. Remote Sens., vol. GE-25, no.5, pp.623-628, September, 1987.
A theory is described to relate the statistical properties of the fields backscattered from rough surfaces to those of speckle in synthetic aperture radar (SAR) images. The expressions are derived for the autocorrelation and cross-correlation functions of speckle intensity in both single-look and multilook images of stationary random rough surfaces in terms of the SAR system parameters and the autocorrelation function of backscatter radar cross-section fluctuations. It is shown that if the correlation scale of cross-section fluctuations is comparable with or greater than the SAR resolution, the correlation functions of speckle intensity depend on those of the cross-section fluctuations. This property, therefore, may be applied to image classification. Comparison of the theory with computer simulation shows good agreement.
"Multilook images of ocean waves by synthetic aperture radars," IEEE Trans. Antennas Propagat., vol. AP-35, no.3, pp.313-317, March, 1987.
A property of multilook processing of synthetic aperture radar (SAR) data is that a time lapse exists between subapertures, so that they contain information about a scattering surface at different times. Reported here is a theoretical study on the images of dynamic ocean waves processed by this technique. It is shown that due to the time lapse the subimages of a moving ocean wave differ in position depending on the look number and the wave phase velocity. Such images cannot be enhanced by the incoherent addition so much as those of stationary surfaces. The difference in image position can be corrected by defocusing the azimuth reference signal by the same amount as for the correction of defocusing induced by the wave motion. Discussions are presented on the correction of image positions and on the effect of defocusing. The property of the time lapse could be applied to estimating not only the phase velocity of ocean waves but also temporal changes in general scattering surfaces.
"On the images of ocean waves by synthetic aperture radars," in Wave Propagation and Scattering, Oxford: Oxford University Press, (ed. B.J.Uscinski), pp.297-332, 1986.
The theory of synthetic aperture radar (SAR) imaging of dynamic ocean surface waves is presented. The principles are described of single- and muli-look processing of SAR data from a stationary and moving point target with special emphasis on the motion effects inherent to multi-look processing. The theory is extended to the imaging of diffusely scattering ocean waves, where image modulation by backscatter cross sections, velocity bunching and tilt or range bunching is discussed. Image degradation by defocusing and the finite lifetime of scatterers are also considered.
"Multilook processing of synthetic aperture radar data from dynamic ocean surfaces," Pattern Recognition Lett., vol.4, no.4, pp.305-314, September, 1986.
The basic theory is described of multi-look processing of synthetic aperture radar (SAR) data from dynamic ocean surfaces. The principal effects inherent to multi-look SAR on the impulse response function are first illustrated using a moving point target model. The theory is then extended to the multi-look imagery of diffusely scattering ocean surface waves. Finally, the use of a sliding synthetic aperture is suggested to investigate the short life-times of small scale surface waves that are considered to be the predominant scatterers on the ocean surface.
"On the multilook images of moving targets by synthetic aperture radars," IEEE Trans. Antennas Propagat., vol. AP-33, no.8, pp.823-827, August, 1985.
In the multilook processing of synthetic aperture radar (SAR) data, subapertures are synthesized at different center times so that a time-lapse exists between looks. This does not affect the imaging of stationary targets but if targets are in motion, the information content about the targets differs from look to look. The purpose of this paper is to investigate the effects of motions on the SAR multilook images of moving targets. Expressions are derived for the impulse response function from a moving point target in terms of the look number and the nature of the motion. Discussions are given only on the effects inherent to multilook processing.
"Two-dimensional imaging mechanisms of ocean waves by synthetic aperture radars," J. Phys., D : Appl. Phys., vol.17, no.1, pp.25-42, January 14, 1984.
Expressions are derived and discussed for the local mean intensity of the two-dimensional images of ocean waves produced by a synthetic aperture radar (SAR). Accounts are taken of the spatial and temporal changes of both back-scattered radar cross-section and ocean wave height. It is shown that the images of ocean waves are distorted as a result of radar layover. This imaging process is similar to velocity bunching but the effect is largest for range waves and it vanishes for azimuth waves. The mechanism of defocusing is also investigated. If reference signals are designed for stationary objects, the images of dynamic waves are always defocused, irrespective of their propagation direction and of the types of imaging processes, including radar layover, velocity bunching and back-scattered cross-section. Defocusing originates from the systematic degradation and upgradation of point spread (impulse response) functions associated with the periodic structure of the waves. The images can be enhanced by applying a defocused azimuth reference signal and the amount of defocusing for optimum images depends only on the wave phase velocity and propagation direction.
"Effect of defocusing on the images of ocean waves produced by synthetic aperture radars," in Satellite Microwave Remote Sensing, ( Chichester: Ellis Horwood, ed. T.D.Allan ), pp.209-234, 1983.
The effects of defocusing on the images of dynamic sea surfaces produced by SAR are of considerable current interest (Shuchman & Zelenka 1978, Jain 1978, Valenzuela 1980, Alpers & Rufenach 1979, Raney 1981). Two main theories have been postulated to explain the experimental observations in which the images of ocean waves are enhanced by adjusting the azimuth focus of a SAR processor. One of them is that the azimuth component of the wave phase velocity changes the relative velocity of the radar platform (Shuchman & Zelenka 1978, Jain 1978, Valenzuela 1980), and the other is that defocusing is caused by the slant-range component of the acceleration associated with the wave orbital motion (Alpers & Rufenach 1979). In this chapter, we investigate the defocusing effect by applying a different approach to the formulation of SAR imagery of dynamic objects from the conventional phase perturbation method. The main advantage of the present approach is that it is rigorous and could lead to simple analytical results (Rufenach & Alpers 1981). We consider that the main contribution to the final image is backscattered amplitude (power) modulation arising from surface roughness and/or surface tilt (Elach & Brown 1977), and the velocity bunching effect (Valenzuela 1980, Alpers & Rufenach 1979, Rufenach & Alpers 1981, Ouchi 1982a) is not included. The cases where the two effects contribute to the images of azimuth waves have been previously discussed by Ouchi (1982b). We also consider the spatial and temporal random fluctuations of backscattered amplitude which in effect degrade image quality (Ouchi 1982c, Ouchi 1981, Raney 1980, Raney & Shuchman 1978, Raney 1981). We then derive and discuss expressions for the local mean intensity and constrast of the images of ocean waves propagating in an arbitrary direction. It is shown that if the reference signal of the SAR processor is matched to a stationary point target, the images of waves are always defocused, irrespective of their propagation direction. The images can be enhanced by applying a defocused reference signal, and the amount of defocus depends on the phase velocity and propagation direction of ocean waves.
"Defocus dependence on ocean wave shape in synthetic aperture radar imagery," Opt. Quant. Electron., vol.15, no.4, pp.355-357, July, 1983.
In a previous paper [1] we examined the effect of defocusing on the images of sinusoidal ocean waves produced by synthetic aperture radar (SAR) and suggested that defocusing may depend on the shape of waves. In this note, we investigate this dependence in a little more detail. For simplicity, the image structure in the azimuth (along-track) direction only will be considered.
"Imagery of ocean waves by synthetic aperture radars," Apl. Phys., vol. B-29, no.1, pp.1-11, January, 1982.
We consider the image structure of ocean waves produced by a synthetic aperture radar (SAR). The theory takes account of the spatial and temporal dependence of small scale capillary waves and large scale waves. Detailed discussions are given to the imaging process in which both amplitude modulation and velocity bunching contribute to the final images. Discussions are also presented on the effect of defocusing on the images of azimuth waves.
"Effect of random motion on synthetic aperture radar imagery," Opt. Quant. Electron., vol.14, no.3, pp.263-275, May, 1982.
We consider the effect of the random motion of capillary waves on the mean intensity and the contrast of the images of dynamic ocean surfaces, with such features as surface waves and current boundaries, produced by a synthetic aperture radar (SAR).
"Statistics of speckle in synthetic aperture radar imagery from targets in random motion," Opt. Quant. Electron., vol.13, no.2, pp.165-173, March, 1981.
We derive and discuss expressions for the autocorrelation functions of speckle patterns in synthetic aperture radar imagery both when the scattering surface is stationary and in random motion. The effects of the motion of a point target on the poin spread function are also discussed.
"First-order statistics of speckle from hard-edged apertures," Ph.D. Thesis, University of London, 1980.
The present thesis is concerned with the first-order statistics and contrast of speckle patterns in and near the far-field and image plane, formed by optical systems having hard-edged pupils. We start, in Chapter one, by reviewing some previous work on the subject. In Chapter two, we give a detailed analysis for the first-order statistics of speckle patterns in and near the far-field. The theory involves assumptions about the fisrt-order probability density function and autocorrelation function of the surface height of scattering media and these are assumed to have Gaussian forms. By applying small-angle scaler diffraction theory, we derive expressions for the constrast in terms of the statistical parameters of rough surfaces and the geometry of the optical system. It is shown that the contrast of speckle patterns produced by hard-e