With the rapid development of maritime transportation, marine environmental pollution has become more and more serious. In particular, oil leakage from ships is regarded as a major threat to marine environmental pollution. Synthetic Aperture Radar (SAR) imagery has become an important technology for marine environment monitoring because of its high resolution, wide coverage, and less influence by light and weather conditions. However, SAR images cannot provide realtime information about ships. The Automatic Identification System (AIS) can provide dynamic information about ships, such as real-time position, heading, speed, etc., helping to identify ships in the sea area. This study combined SAR imagery and AIS data to detect oil spills at sea and identify suspicious oil-discharging vessels. First, an improved U-Net model was adopted to detect oil spills and ships in SAR imagery. To achieve high detection performance, the U-Net was modified by using a lightweight MobileNetv3 backbone architecture, convolutional block attention module (CBAM), atrous spatial pyramid pooling (ASPP), and full-scale feature aggregation. Experimental results showed that the proposed U-Net model improved the detection accuracy of oil spills and reduced the misclassification between oil spills and look-alikes. Then, the AIS data corresponding to the SAR image was collected and the trajectories of ships passing near the SAR acquisition time can be screened out. The study compared AIS data with SAR detection results to look for the ship that was closest to the oil spill and whose navigation trajectory was almost parallel to the direction of the oil spill extension. Thus, it can be inferred that the ship was suspected of discharging oil pollution. Through experiments on oil pollution incidents, the effectiveness of combining SAR and AIS in tracking oil-discharging ships was verified.
The fully polarimetric synthetic aperture radar (PolSAR) with high resolution and quad-polarization data has shown the ability to provide better interpretation and high classification accuracy. But PolSAR image quality is critical disturbed due to speckle noise not only in the three intensities (HH, HV, VV) of PolSAR but also in the three complex correlation terms (HH-VV, VV-HV and HV-HH). Thus, reducing speckle noise level while preserving polarization scattering mechanisms and spatial resolution of PolSAR becomes a challenging task. Several filters widely used in SAR images can achieve effective speckle reduction, but can also lead to edge blurring and strong reflective scatter depressing. To mitigate these deficiencies, in the study, we proposed an improved speckle filtering approach by combining existing filters (such as Lee or Sigma filters) with edge/strong target detectors. Before performing filtering, strong targets and edges are detected and preserved by the proposed detectors. By the detection preprocessing, the improved filtering approach can not only reduce speckle noise but also preserve the target signature. The effectiveness of the proposed improved speckle filters is validated by two kinds of PolSAR data, one obtained by NASA Airborne Synthetic Aperture Radar System (AIRSAR) and the other data obtained by Advanced Land Observing Satellite(ALOS). From experimental results, the proposed improved filter provides promising results for suppressing speckle noise and preserving the potential targets.
The fully polarimetric SAR (PolSAR) data offers four polarimetric modes (i.e., HH, HV, VH, VV) and have shown the ability to provide better interpretation than single polarization case, which led to high classification accuracy. In this study, an efficient classification method for PolSAR data based on a subspace projection segmentation (SPS) approach is proposed to improve the classification accuracy. Instead of performing the comparison of multi-dimensional (MD) polarimetric feature vectors, the SPS first transforms the MD polarization feature vectors into one-dimensional (1D) projection lengths by projecting the feature vectors onto one reference subspace which is chosen to maximize the separation of two types of data. After the transformation, any 1D thresholding technique, such as the Otsu’s thresholding, can be applied to perform segmentation efficiently, which results in the reduction of computation complexity in segmentation. The proposed SPS can divide the data into proper homogeneous regions, that is, PolSAR data with similar polarization features being grouped together into regions. In the study, the polarimetric feature vectors are extracted from the coherent/covariance matrices obtained by the polarimetric scattering information. In addition, the referenced projection subspace is selected based on the coherent/covariance matrices of PolSAR data. Finally, the performance of the proposed SPS method is validated by simulations on PolSAR data obtained by NASA Airborne Synthetic Aperture Radar System (AIRSAR) during the PacRim II project and Advanced Land Observing Satellite (ALOS). Simulation results show that the proposed approach can reduce the computational complexity more effectively than other existing methods, and also achieve good classification accuracy.
In the study, we develop a multiple constrained signal subspace projection (SSP) approach to target detection. Instead of using single constraint on target detection, we design an optimal filter with multiple constraints on desired targets by using SSP. The proposed SSP approach fully exploits the orthogonal property of two orthogonal subspaces: one denoted signal subspace containing desired and undesired/background targets; the other denoted noise subspace, which is orthogonal to signal subspace. By projecting the weights of the detection filter on the signal subspace, the proposed SSP can reduces some estimation errors in target signatures and alleviate the performance degradation caused by uncertainty of target signature. The SSP approach can detect desired targets, suppress undesired targets and minimize the interference effects. In experiments, we provide three methods in selecting multiple constraints of the desired target: Kmeans, principal eigenvectors and endmenber extracting techniques. Simulation results show that the proposed SSP with multiple constraints selected by K-means has better detection performance. Furthermore, the proposed SSP with multiple constraints is a robust detection approach which could overcome the uncertainty of desired target signature in real image data.
In the study, we proposed an adaptive filter with multiple constrains based on the generalized sidelobe canceller (GSC) structure for target detection of hyperspectral images. The proposed filtering approach can alleviate the performance degradation in target detection caused by estimation errors in spectral signature of the desired target or some random noise by unknown interference. First, we design an optimal filter to minimize the interference effect with multiple constrains including unit gain response on desired target and null response on undesired targets. The optimal filter can detect the desired target, suppress the undesired targets and minimize the interference effect. Next, an adaptive filter with GSC structure is proposed to transform the constrained minimization problem into an equivalent unconstrained minimization. The structure of GSC contains two branches: the upper branch is a filter with fixed weights wf designed by multiple constrains to reserve the desired target and interference; the lower branch contains a blocking matrix B and an adaptive filter with weights wa. Matrix B blocks the desired target and preserve the interference. The adaptive filter can be designed to minimize the interference effect without constrains. Simulations validate the effectiveness of the proposed adaptive filter with GSC structure which is robust to the random errors in spectral signature of the desired target.
KEYWORDS: Orthogonal frequency division multiplexing, Matrices, Antennas, Bismuth, Telecommunications, Monte Carlo methods, Algorithm development, Computer simulations, Data communications, Error analysis
In the study, we propose an efficient subspace-based semiblind channel estimation for multiple-input–multiple-output (MIMO) space–time code (STC) orthogonal frequency-division multiplexing (OFDM) systems. We first proposed a forward-backward estimation (FBE) method which can improve the channel estimation accuracy by using both the forward and backward receiving data. Then, based on the symmetric property of the forward and backward smoothed correlation matrix, we develop a fast forward-backward (FFB) estimation method which estimates the noise subspace by performing eigen-decomposition of two half dimensionality sub-matrices obtained from the forward and backward smoothed correlation matrix. FFB achieves the same performance as the FBE but only requires one-fourth computation complexity of FBE. Computer simulations demonstrate the effectiveness and accuracy in channel estimation of the proposed FFB for the MIMO STC-OFDM systems.
KEYWORDS: Image classification, Hyperspectral imaging, Image filtering, Electronic filtering, Signal detection, Signal to noise ratio, Monte Carlo methods, Interference (communication), Image processing, Library classification systems
In this study, we propose an efficient classification which combines signal subspace projection (SSP) and partial filtering technique for hyperspectral images. To reduce the computation complexity in image classification, we exploit high degree correlations in spectral and spatial domains. During training process, image bands are first partitioned into several groups for each desired class by Maximum Correlation Band Clustering (MCBC) approach. Then, we design partial filters for each band group by SSP approach. Finally, the SSP-based partial filtering (SSPPF) are combined using corresponding weights for each class. For real image classification, simulations validate the proposed SSPPF can achieve the performance of SSP with less computation complexity. Generally, the proposed method requires only 1/ k2 computations of SSP, if image is partitioned into k groups.
In hyperspectral imagery, greedy modular eigenspace (GME) was developed by clustering highly correlated bands into a smaller subset based on the greedy algorithm. Unfortunately, GME is hard to find the optimal set by greedy scheme except by exhaustive iteration. The long execution time has been the major drawback in practice. Accordingly, finding the optimal (or near-optimal) solution is very expensive. Instead of adopting the band-subset-selection paradigm underlying this approach, we introduce a simulated annealing band selection (SABS) approach, which takes sets of non-correlated bands for high-dimensional remote sensing images based on a heuristic optimization algorithm, to overcome this disadvantage. It utilizes the inherent separability of different classes embedded in high-dimensional data sets to reduce dimensionality and formulate the optimal or near-optimal GME feature. Our proposed SABS scheme has a number of merits. Unlike traditional principal component analysis, it avoids the bias problems that arise from transforming the information into linear combinations of bands. SABS can not only speed up the procedure to simultaneously select the most significant features according to the simulated annealing optimization scheme to find GME sets, but also further extend the convergence abilities in the solution space based on simulated annealing method to reach the global optimal or near-optimal solution and escape from local minima. The effectiveness of the proposed SABS is evaluated by NASA MODIS/ASTER (MASTER) airborne simulator data sets and airborne synthetic aperture radar images for land cover classification during the Pacrim II campaign. The performance of our proposed SABS is validated by supervised k-nearest neighbor classifier. The experimental results show that SABS is an effective technique of band subset selection and can be used as an alternative to the existing dimensionality reduction method.
Satellite remote sensing images can be interpreted to provide important information of large-scale natural resources, such
as lands, oceans, mountains, rivers, forests and minerals for Earth observations. Recent advances of remote sensing
technologies have improved the availability of satellite imagery in a wide range of applications including high
dimensional remote sensing data sets (e.g. high spectral and high spatial resolution images). The information of high
dimensional remote sensing images obtained by state-of-the-art sensor technologies can be identified more accurately
than images acquired by conventional remote sensing techniques. However, due to its large volume of image data, it
requires a huge amount of storages and computing time. In response, the computational complexity of data processing
for high dimensional remote sensing data analysis will increase. Consequently, this paper proposes a novel classification
algorithm based on semi-matroid structure, known as the parallel k-dimensional tree semi-matroid (PKTSM)
classification, which adopts a new hybrid parallel approach to deal with high dimensional data sets. It is implemented by
combining the message passing interface (MPI) library, the open multi-processing (OpenMP) application programming
interface and the compute unified device architecture (CUDA) of graphics processing units (GPU) in a hybrid mode. The
effectiveness of the proposed PKTSM is evaluated by using MODIS/ASTER airborne simulator (MASTER) images and
airborne synthetic aperture radar (AIRSAR) images for land cover classification during the Pacrim II campaign. The
experimental results demonstrated that the proposed hybrid PKTSM can significantly improve the performance in terms
of both computational speed-up and classification accuracy.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.