Urban land use classification using synthetic aperture radar

Abstract

This paper presents several approaches to the use of radar imagery for land use classification of urban and near-urban areas. The use of L(HH) (L band, horizontal transmit and horizontal receive) data is emphasized because it is these types of data obtained by Seasat-A (and in November 1981 by Shuttle radar) which are most generally available. For urban area studies using imaging radar the effect of processing in an off-zero doppler (‘squint’) mode, the presence of large diffuse scatters and the possibility of height measurements are discussed. Each approach provides information and also requires supporting ground truth which are unique to radar remote sensing. For some areas the coupling of data from the microwave portion of the spectrum to the data available in the visible and near visible realms may improve the classification of urban and near-urban land use. However, the radar data are not without their own limitations which may be imposed by either the system or the nature of the imaged scene. A proper knowledge of these limitations can permit us to turn a perceived defect into a decided advantage. The metropolitan area of Los Angeles provides the geographic background for this study.     

Formation flying orbit design for the distributed synthetic aperture radar satellite

Formation flying orbit design is one of the key technologies for system design and performance analysis of the distributed SAR satellites. The approximately analytic solution of the passive stable formation flying orbit elements is explored based on the expansion form of Kepler's equation. A new method of orbital parameters design for three-dimensional formation flying SAR satellites is presented, and the precision of the orbital elements is analyzed. Formation flying orbit elements are calculated for the L-Band distributed SAR satellites using the formulas deduced in this paper. The accuracy of the orbital elements is validated by the computer simulation results presented in this paper.     

Suppression of speckle in synthetic aperture radar images using wavelet

A filter for suppressing speckle in synthetic aperture radar (SAR) images utilizing wavelet is proposed. The filter suppresses speckle by reducing the amplitude of the detail images in wavelet subspaces, while preserving edges by releasing the amplitude reduction around edges; information on edges, contained in the detail images, is utilized for edge detection. Simulations and application to SAR images have shown that the performance of the filter is satisfactory in both smoothing and edge preservation, and in generating visually-natural images as well.     

Gini index-based land-cover classification using polarimetric synthetic aperture radar

Classification of the Earth's surface types is one of the important remote-sensing applications of radar polarimetry. An unsupervised classification scheme based on the use of entropy and alpha angle is widely used for land-cover classification using multi-polarization radar images. The polarimetric entropy and the alpha angle are used to characterize a target's randomness and scattering mechanism, respectively. Here, we replace the entropy by the Gini index. Evaluation of the Gini index is computationally efficient. It also overcomes the drawback encountered in entropy evaluation, namely, the use of logarithmic operation. We develop and validate an unsupervised classification technique based on the use of the Gini index and alpha angle and show that it performs better than the classic entropy/alpha classification technique. We have also used the Gini/alpha method with anisotropy and complex Wishart distribution to design a complete land-cover classification scheme. The proposed classification scheme performs better than the entropy/alpha land-cover classification scheme.     

A technique to estimate topsoil thickness in arid and semi-arid areas of north-eastern Jordan using synthetic aperture radar data

A Japanese Earth Resources Satellite (JERS)-1 L-band synthetic aperture radar (SAR) dataset was used for estimating topsoil thickness, of different types, in arid and semi-arid areas in north-eastern Jordan. In this research, the relationship between remotely sensed data, backscattering coefficient and the thickness of topsoil was investigated. Based on the dielectric constant properties of the topsoil samples, the relationship between the backscattering coefficient and the topsoil thickness was obtained by developing a multilayer modelling analysis. Using this model, the topsoil thickness had been estimated by means of the derived backscattering coefficients from JERS-1 SAR image. The estimated thickness values of the different topsoil types were found to be comparable with field ground data. The estimated minimum thickness for hardpan topsoil is 55?cm; Qaa topsoil, 74?cm; and topsoil of herbaceous area, 46?cm, while the estimated maximum thickness is more than 98?cm, more than 100?cm, and 82?cm, respectively. Ground data, on the other hand, revealed the minimum thickness for hardpan topsoil to be 50?cm; Qaa topsoil, 70?cm; and topsoil of herbaceous area, 40?cm, while the maximum thickness is more than 120?cm, more than 100?cm, and 80?cm, respectively.     

Change detection using distance-based algorithms between synthetic aperture radar polarimetric decompositions

The present study introduces distance based change detection (CD) algorithms in polarimetric synthetic aperture radar (PolSAR) data. PolSAR images, due to interactions between electromagnetic waves and target and because of the high spatial resolution, can be used to study changes in the Earth’s surface. The purpose of this paper is to use features extracted from the fully-polarimetry imaging radar that involved Yamaguchi four-component and H/A/α decomposition based on the distance between the vectors of features for CD. We first extract features from polarimetric decompositions of multi-looked covariance (or coherency) matrix data. We then use two well-known distance measures namely Canberra and Euclidean distances for measuring the similarity between the vectors of polarimetric decompositions at different times. Assessment of incorporated methods is performed using different criteria, such as overall accuracy, area under the receiver operating characteristic curve, and false alarms rate. The results of the experiments show that Canberra distance has better performance with high overall accuracy and low false alarm rate than Euclidean distance and other compared algorithms to detect changes.     

Monitoring dewatering induced subsidence and fault reactivation using interferometric synthetic aperture radar

Using multitemporal differential interferometric synthetic aperture radar analysis integrated with pumping and site geologic data we present evidence for hydrologically induced large subsidence in and around an ongoing open‐pit mine with intensive dewatering operations. Analysis of numerous differential synthetic aperture radar interferometry (DInSAR) pairs spanning the period 1993 to 2001 reveals the abrupt appearance of these features to intervals of a few to several months. Along a section through the anomaly, we plotted dewatering associated changes in the groundwater levels at monitoring wells. We also used DInSAR to extract several individual kilometre‐lengths, centimetre amplitude normal fault reactivation events in the alluvial sediments adjacent to the mine dewatering operation. High‐resolution remote sensing analyses provide strong evidence that these features align with faults active in the last several thousand years. We interpret these reactivations as mechanically involving only the upper few hundred metres of the existing fault plane above the alluvial aquifer affected by the mine dewatering.     

Noise reduction in synthetic aperture radar images using fuzzy logic and genetic algorithm

Synthetic aperture radar (SAR) is a self-illuminating imaging technique; it produces high resolution images in all weather conditions, day and night. SAR images are widely accepted and used by many application scientists. However, the SAR images are corrupted with speckle noise. Speckle noises are caused by random interference of electromagnetic signals scattered by the object surface within one resolution element. The amount of noise and distribution of noise corrupting the image is unpredictable. Conventional noise filters are quantitative in nature; they are not well suited for uncertainty problems. Fuzzy logic is capable of handling uncertainty. In this work, noisy pixels in the images are identified by using fuzzy rules and filtered using fuzzy weighted mean, keeping the healthy pixels unchanged. The optimum value of parameters used in defining fuzzy membership function is determined by using genetic algorithm (GA). Reducing noise and simultaneously preserving image details are the two most desirable characteristics of noise filters. Peak signal-to-noise ratio (PSNR) and edge preserving factor (EPF) are used to evaluate the performance of the proposed fuzzy filter. SAR images affected by varying amounts of speckle noise are used to evaluate the performance. It was observed that the proposed filter suppresses noise and preserves image edges.

     

Mapping arctic landfast ice extent using L-band synthetic aperture radar interferometry

In recent years methods have been developed to extract the seaward landfast ice edge from series of remote sensing images, with most of them relying on incoherent change detection in optical, infrared, or radar amplitude imagery. While such approaches provide valuable results, some still lack the required level of robustness and all lack the ability to fully automate the detection and mapping of landfast ice over large areas and long time spans. This paper introduces an alternative approach to mapping landfast ice extent that is based on coherent processing of interferometric L-band Synthetic Aperture Radar (SAR) data. The approach is based on a combined interpretation of interferometric phase pattern and interferometric coherence images to extract the extent and stability of landfast ice. Due to the low complexity of the base imagery used for landfast ice extraction, significant improvements in automation and reduction of required manual interactions by operators can be achieved. A performance analysis shows that L-band interferometric SAR (InSAR) data enable the mapping of landfast ice with high robustness and accuracy for a wide range of environmental conditions.     

Environmental monitoring of natural disasters using synthetic aperture radar image multi-directional characteristics

Synthetic aperture radar (SAR) imaging can penetrate rain, snow, fog, and mist, providing an effective precipitation detector under a wide range of atmospheric conditions. This powerful technique operates under any weather conditions and can detect environmental changes or help in the evaluation of natural disasters. Multi-temporal change detection is important for monitoring disasters, but commonly applied wavelet transforms are not ideal for capturing change information. This paper presents a new multi-directional change detection (MDCD) method designed to improve the change detection accuracy using multiple SAR images. This method employs double-density dual-tree complex wavelet transforms (DDDT-CWT) theory, which allows the capture of multi-directional information. The MDCD method can provide information for 16 directions at any decomposed scale, which allows for change detection in multiple SAR images collected over time. We used the MDCD algorithm method to analyse SAR images from actual natural disasters and successfully identified environmental changes over time using SAR images.     

Agricultural crop condition monitoring using airborne C-band synthetic aperture radar in southern Alberta

Applications of airborne C-band synthetic aperture radar imagery for determining variations in agricultural crop characteristics were investigated at a test site in southern Alberta, Canada. Synthetic aperture radar (SAR) imagery and ground-based crop characteristics data were acquired on 19–20 July 1994 for wheat, canola, beans, peas, and wheat + alfalfa cultivated under a variety of irrigation conditions. The results indicate that the statistically significant relationships that were derived between the ground-based data and SAR imagery are a function of crop type, crop condition parameter, and image processing procedures, and that crop characteristics such as leaf area index and plant height are negatively correlated with radar backscatter.     

Multiregion level-set partitioning of synthetic aperture radar images

The purpose of this study is to investigate synthetic aperture radar (SAR) image segmentation into a given but arbitrary number of gamma homogeneous regions via active contours and level sets. The segmentation of SAR images is a difficult problem due to the presence of speckle which can be modeled as strong, multiplicative noise. The proposed algorithm consists of evolving simple closed planar curves within an explicit correspondence between the interiors of curves and regions of segmentation to minimize a criterion containing a term of conformity of data to a speckle model of noise and a term of regularization. Results are shown on both synthetic and real images.     

Statistical investigations of the synthetic aperture radar images

The focused and non-focused image-forming methods for airborne synthetic aperture radar (SAR) are investigated. The more precise receiving signal model which includes not only traditional additive noise but the multiplicative noise which is the consequence of the aircraft's velocity and height fluctuations is developed. Results of numerical calculations of the root-mean-square (rms) errors of the set of principal image parameters as functions of the additive noise variance and multiplicative noise standard deviation for several weighting windows and the histograms of the distributions of some image parameters are presented.     

Ocean sand ridge signatures in the Bohai Sea observed by satellite ocean color and synthetic aperture radar measurements

Satellite measurements from Synthetic Aperture Radar (SAR) and the Moderate Resolution Imaging Spectroradiometer (MODIS) onboard the Aqua platform are used to study the ocean sand ridges in the eastern Bohai Sea in China. Even though the imaging mechanisms for SAR and MODIS-Aqua remote sensing are different, the sand ridges are shown to have exactly the same patterns in images from both sensors. Therefore, the location, extension and coverage of the ocean sand ridges can be detected and cross-examined by both SAR and MODIS-Aqua observations. Satellite images show quite different sand ridge distribution pattern from the published bathymetry map (based on in situ data) that shows six sand ridges in the area. 10 finger-shaped sand ridges are identified from satellite observations. The tidal-current/sand-ridge interaction driven physical and optical changes are assessed and evaluated. The existence of sand ridges causes enhanced water diffuse attenuation coefficient K_d(490) and elevated normalized water-leaving radiance at the red and near-infrared (NIR) wavelengths. The sea surface over the sand ridges experiences significant seasonal variability of water turbidity and shows remarkable differences from nearby ocean regions. During winter, K_d(490) values are about 2-3 m^− 1 over the ridges, while the maximum K_d(490) in the neighboring oceans is approximately 1.5 m^− 1. In summer, the enhancement of the sea surface turbidity is less significant than that which occurs in winter.     

Range ambiguity suppression in a synthetic aperture radar using pulse phase coding and two-pulse cancellation

ABSTRACT

In the presence of range ambiguity, the synthetic aperture radar (SAR) systems suffer from image aliasing, which dramatically degrades the quality of SAR images. In this article, an easy-to-implement technique for range ambiguity suppression is investigated, which is based on phase coding in the transmit pulse dimension, referred to as pulse phase coding (PPC). By properly designing the PPC series, it is possible to discriminate the range ambiguous echoes from different range areas in the Doppler frequency domain. To further suppress the range ambiguous echo, a two-pulse cancellation (TPC)-based SAR imaging method is proposed, which improves the quality of SAR images in the presence of range ambiguity. The proposed two-pulse cancellation is performed followed by the azimuth compression. The simulation results demonstrate the effectiveness of the proposed method.     

Nearshore circulation and synthetic aperture radar: An exploratory study

We use a sequence of synthetic aperture radar (SAR) images to map differences in the flood and ebb tidal currents in a cove along the coast of Nova Scotia, Canada. The asymmetry in the tidal flow determines the flushing rate of the cove which, in turn, has a significant effect on biological production within the cove and its potential for aquaculture. We find significant differences in the SAR images collected on flood and ebb tides. Specifically there are well-defined lines on the flood images and large whorls on the ebb. A three-dimensional hydrodynamic model of the tidal currents is used to interpret the SAR images. In particular we use the model flow fields to wind back the SAR images to an earlier stage of the tide in an attempt to determine the physical origin of the features in the images. We conclude that the most likely explanation for the ebb-tide whorls is the advection of surface slicks.     

Influence of incidence angle on detecting flooded forests using C-HH synthetic aperture radar data

Hydrology is the single most important abiotic factor in the formation and functioning of a wetland. Many limitations still exist to accurately characterizing wetland hydrology over large spatial extents, especially in forested wetlands. Imaging radar has emerged as a viable tool for wetland flood mapping, although the limitations of radar data remain uncertain. The influence of incidence angle on the ability to detect flooding in different forest types was examined using C-HH Radarsat-1 data (23.5°, 27.5°, 33.5°, 39.0°, 43.5°, and 47.0°) during the leaf-off and leaf-on seasons. The ability to detect flooding under leaf-on conditions varied much more according to incidence angle while forest type (open canopy tupelo-cypress, tupelo-cypress, and bottomland hardwood) had a greater effect during the leaf-off season. When all forest types were considered together, backscatter generally decreased with increasing incidence angle under all conditions (2.45 dB between 23.5° and 47.0° flooded, leaf-off; 2.28 dB between 23.5° and 47.0° not flooded, leaf-off; 0.62 between 23.5° and 43.5° flooded, leaf-on; 1.73 dB between 23.5° and 43.5° not flooded, leaf-on; slope was not constant between incidence angles), but the distinction between flooded and non-flooded areas did not decline sharply with incidence angle. Differentiation of flooded and non-flooded forests was similar during the leaf-off and leaf-on seasons. The ability to detect inundation under forest canopies was less than expected at smaller incidence angles and greater than expected at larger incidence angles, based on the results of previous studies. Use of a wider range of incidence angles during the entire year increases the temporal resolution of imagery which may, in turn, enhance mapping of inundation beneath forest canopies.     

Polarimetric synthetic aperture radar image segmentation by convolutional neural network using graphical processing units

Image segmentation is an important application of polarimetric synthetic aperture radar. This study aimed to create an 11-layer deep convolutional neural network for this task. The Pauli decomposition formed the RGB image and was used as the input. We created an 11-layer convolutional neural network (CNN). L-band data over the San Francisco bay area and C-band data over Flevoland area were employed as the dataset. For the San Francisco bay PSAR image, our method achieved an overall accuracy of 97.32%, which was at least 2% superior to four state-of-the-art approaches. We provided the confusion matrix over test area, and the kernel visualization. We compared the max pooling and average pooling. We validated by experiment that four convolution layers perform the best. Besides, our method gave better results than AlexNet. The GPU yields a 173× acceleration on the training samples, and a 181× acceleration on the test samples, compared to standard CPU. For the Flevoland PSAR image, our 11-layer CNN also gives better overall accuracy than five state-of-the-art approaches. The convolutional neural network is better than traditional classifiers and is effective in remote sensing image segmentation.     

对不同工作模式星载合成孔径雷达的侦察研究

条带模式和聚束模式是合成孔径雷达常用的两种工作模式,尤其是聚束模式的精度高,在军事上发挥着重要的作用。在研究两种工作模式原理的基础上,根据合成孔径雷达的特点,展开了对合成孔径雷达工作模式的侦察研究,仿真实验验证了侦察的可行性。