Proposed Processing Chain Research Articles

Deep Learning-Based Link Configuration for Radar-Aided Multiuser mmWave Vehicle-to-Infrastructure Communication

Configuring millimeter wave links following a conventional beam training protocol, as the one proposed in the current cellular standard, introduces a large communication overhead, especially relevant in vehicular systems, where the channels are highly dynamic. In this paper, we propose the use of a passive radar array to sense automotive radar transmissions coming from multiple vehicles on the road, and a radar processing chain that provides information about a reduced set of candidate beams for the links between the road-infrastructure and each one of the vehicles. This prior information can be later leveraged by the beam training protocol to significantly reduce overhead. The radar processing chain estimates both the timing and chirp rates of the radar signals, isolates the individual signals by filtering out interfering radar chirps, and estimates the spatial covariance of each individual radar transmission. Then, a deep network is used to translate features of these radar spatial covariances into features of the communication spatial covariances, by learning the intricate mapping between radar and communication channels, in both line-of-sight and non-line-of-sight settings. The communication rates and outage probabilities of this approach are compared against exhaustive search and pure radar-aided beam training methods (without deep learning-based mapping), and evaluated on multi-user channels simulated by ray tracing. Results show that: (i) the proposed processing chain can reliably isolate the spatial covariances for individual radars, and (ii) the radar-to-communications translation strategy based on deep learning provides a significant improvement over pure radar-aided methods in both LOS and NLOS channels.

Ultra-High-Resolution Spaceborne and Squint SAR Imaging for Height-Variant Geometry Using Polynomial Range Model

In this article, we propose an entire processing chain that addresses the geometrical defocusing problems caused by a large orbital arc, a high squint, and topography variations in ultra-high-resolution (UHR) synthetic aperture radar (SAR) operations. Slant-range histories are formulated using fifth-order polynomials to accurately express radar signals with a UHR spaceborne and squint SAR geometry. SAR focusing is first conducted with a constant height assumption. Thereafter, residual topography errors are compensated. Several beneficial contributions are included in the proposed processing chain. First, to achieve constant-height SAR focusing, we improve the typical least-square decomposition-based Stolt interpolations in terms of the image quality and computation time. Second, we clarify how illuminated areas are mapped in a two-dimensional imaging domain after polynomial-based focusing, which is completely different from the well-known relationship between the closest range and zero-Doppler time. Third, a generalized blockwise postfocusing method is devised to correct the range and azimuth defocusing effects induced by height variations. These contributions are verified using numerical simulations. Thus, we can conclude that the proposed method can provide focused SAR images for a height-variant geometry in UHR spaceborne and squint operations.

Airborne Elevation DBF-TOPS SAR/InSAR Method Based on LOS Motion Compensation and Channel Error Equalization

Digital beamforming (DBF) TOPS SAR in elevation is a new synthetic aperture radar (SAR) system, which has the advantage of wide swath coverage and a high signal-to-noise ratio (SNR). In this paper, considering the phase preservation demand for interferometric SAR (InSAR) processing, the complete processing chain for DBF-TOPS SAR/InSAR in elevation is proposed with a wide beam angle and channels’ amplitude and phase errors. Firstly, we analyze the airborne motion compensation method along the line-of-sight direction for TOPS SAR with squint angle. Furthermore, for the large-range beam angle of DBF, the sub-swaths division process is presented for the range-dependent radar look angle, and the sub-swaths division criterion is also given in the analytic expression. Then, the relative amplitude and phase errors’ estimation and compensation method between channels is provided in the range frequency domain based on the pivoting filter with coherence weighting, which is convenient for DBF processing and SNR improvement. Finally, the DEMs are generated under different conditions to compare the phase preservation performance. The effectiveness of the proposed processing chain is verified with both simulated data and airborne real DBF-TOPS SAR/InSAR data.

A Novel GIS-Based Approach for Automated Detection of Nearshore Sandbar Morphological Characteristics in Optical Satellite Imagery

Satellite remote sensing is a valuable tool for coastal management, enabling the possibility to repeatedly observe nearshore sandbars. However, a lack of methodological approaches for sandbar detection prevents the wider use of satellite data in sandbar studies. In this paper, a novel fully automated approach to extract nearshore sandbars in high–medium-resolution satellite imagery using a GIS-based algorithm is proposed. The method is composed of a multi-step workflow providing a wide range of data with morphological nearshore characteristics, which include nearshore local relief, extracted sandbars, their crests and shoreline. The proposed processing chain involves a combination of spectral indices, ISODATA unsupervised classification, multi-scale Relative Bathymetric Position Index (RBPI), criteria-based selection operations, spatial statistics and filtering. The algorithm has been tested with 145 dates of PlanetScope and RapidEye imagery using a case study of the complex multiple sandbar system on the Curonian Spit coast, Baltic Sea. The comparison of results against 4 years of in situ bathymetric surveys shows a strong agreement between measured and derived sandbar crest positions (R2 = 0.999 and 0.997) with an average RMSE of 5.8 and 7 m for PlanetScope and RapidEye sensors, respectively. The accuracy of the proposed approach implies its feasibility to study inter-annual and seasonal sandbar behaviour and short-term changes related to high-impact events. Algorithm-provided outputs enable the possibility to evaluate a range of sandbar characteristics such as distance from shoreline, length, width, count or shape at a relevant spatiotemporal scale. The design of the method determines its compatibility with most sandbar morphologies and suitability to other sandy nearshores. Tests of the described technique with Sentinel-2 MSI and Landsat-8 OLI data show that it can be applied to publicly available medium resolution satellite imagery of other sensors.

A Simple Method for Near-Real-Time Monthly Nighttime Light Image Production

Compared with previous nighttime light (NTL) data, the Black Marble product developed by NASA in recent years is the most advanced NTL product, which plays an important role in regional activity monitoring and special events (process) tracking due to its excellent timeliness (near-real-time release) and global data coverage capability. Currently, among the Black Marble products, the VNP46A2 product is not updated in a timely manner and VN46AX is still under development. Thus, VNP46A1, which is released in near real time, has become a key product available to the public, although it is affected by cloud cover and moonlight. To obtain high-quality near-real-time monthly NTL images and reduce the data gaps before VNP46AX is released, a simple processing scheme based on VNP46A1 is proposed in this letter. The results show that the proposed processing chain can effectively filter cloud-polluted pixels and alleviate the interference of moonlight radiation. The near-real-time monthly NTL (NRTMNTL) data generated by the scheme achieve a higher correlation with VNP46A2 and present a better visual purity. The near-real-time data availability and parameter-free input make the proposed scheme more applicable for short-period events tracking.

EEG signals analysis for epileptic seizures detection using polynomial transforms, linear discriminant analysis and support vector machines

Electroencephalogram (EEG) signals are useful in understanding the human brain diseases like epilepsy which is characterized by an enduring predisposition to generate epileptic seizures and by neurologic, cognitive, psychological and social consequences of these conditions. That is why this paper proposed a novel full processing chain of EEG signals analysis for epileptic seizures detection that applied a new alternative approach for EEG rhythms decomposition. EEG signals are decomposed into their different background rhythms using discrete Jacobi polynomial transforms (JPTs) and a 28-dimension feature vector is extracted in frequency and time domains. Due to the high dimension and redundancy of these features, the linear discriminant analysis (LDA) is applied for dimensionality reduction and informative and discriminative low dimension feature vectors are computed and feed as inputs of the support vector machines (SVM) classifiers. The validation of this processing chain is done experimentally using an online available database which consists of five hundred EEG signals. Analysis demonstrated that the EEG rhythms decomposition using JPTs can be implemented as a single-step process and provides more abundant information for class discrimination. In addition, twelve experiments closely related to clinical applications are examined and promising results are achieved with maximum accuracies between 96.25–100 %. Overall, it is found that the proposed processing chain will be useful in providing an accurate and objective scheme for automatic EEG rhythms decomposition and epileptic seizures detection that can be integrated into implantable devices intended to predict the onset of seizures and trigger a focal treatment to block the seizures progression.

Multiple Optical Sensor Fusion for Mineral Mapping of Core Samples.

Geological objects are characterized by a high complexity inherent to a strong compositional variability at all scales and usually unclear class boundaries. Therefore, dedicated processing schemes are required for the analysis of such data for mineralogical mapping. On the other hand, the variety of optical sensing technology reveals different data attributes and therefore multi-sensor approaches are adapted to solve such complicated mapping problems. In this paper, we devise an adapted multi-optical sensor fusion (MOSFus) workflow which takes the geological characteristics into account. The proposed processing chain exhaustively covers all relevant stages, including data acquisition, preprocessing, feature fusion, and mineralogical mapping. The concept includes (i) a spatial feature extraction based on morphological profiles on RGB data with high spatial resolution, (ii) a specific noise reduction applied on the hyperspectral data that assumes mixed sparse and Gaussian contamination, and (iii) a subsequent dimensionality reduction using a sparse and smooth low rank analysis. The feature extraction approach allows one to fuse heterogeneous data at variable resolutions, scales, and spectral ranges and improve classification substantially. The last step of the approach, an SVM classifier, is robust to unbalanced and sparse training sets and is particularly efficient with complex imaging data. We evaluate the performance of the procedure with two different multi-optical sensor datasets. The results demonstrate the superiority of this dedicated approach over common strategies.

Time-lapse optical flow regularization for geophysical complex phenomena monitoring

In this paper, we introduce a framework for the tracking of geophysical complex phenomena via time-lapse images. It includes the regularization of the derived surface motion maps time series. The proposed processing chain addresses five main challenges: undesired camera movement, missing frames, important photometric changes, weak/repetitive texture and model-free dense spatial transformations. In the proposed framework the motion maps time series are obtained via robust pre-processing steps and optical flow computing. The contribution consists of regularizing the resulting velocity and position time series to minimize a temporal closure error in a subsequent stage. This step serves to alleviate the limitations of existing methods in the context of geophysical monitoring. The temporal closure errors are formulated as linear mappings to inverse using signal priors and two formulations are defined along with illustrative cases. Related methods are discussed and extensive experimentation on simulated datasets is carried out to validate the approach and compare between the different proposed formulations and resolution schemes. Experimental results are presented on time series acquired by ground-based cameras used for the monitoring of Alpine glaciers. The algorithm is computationally efficient, even considering the quantity of processed and generated data, and is run in parallel on multiple cores for speed-up.

A Bayesian framework for joint structure and colour based pixel-wise classification of grapevine proximal images

Estimating the spatial variability of basic agronomic parameters at the scale of the plant is of prime importance for the development and monitoring of Precision Agriculture applications. It is all the more crucial in viticulture where intra-plot variabilities are exacerbated. This paper focuses on the description of the structure of the canopy at the plant scale by proximal imaging. A new framework is proposed for the pixel-wise classification of the grapevine canopy into organs at different phenological stages. The proposed processing chain proceeds in four steps: (i) foreground extraction, (ii) pixel-wise feature extraction, (iii) pixel-wise classification and (iv) spatial regularization. Step (i) is based on colour information only. For step (ii), colour is represented using an RGB triplet while texture is captured using the local structure tensor (LST). Two variants are proposed to associate colour and LST information into a single Euclidean vector. Step (iii) is a Bayesian decision process based on the joint modelling of colour and texture using multivariate Gaussian distributions. Finally, step (iv) combines stochastic relaxation and morphological filtering, allowing for the spatial regularisation of the classification output. This processing chain is applied to the pixel-wise classification of proximal images into grapevine organs. Images were taken from two 0.2 ha plots planted with the red variety “Merlot Noir” in Bordeaux area. Images were taken from an embedded acquisition system at three key phenological stages: flowerhood falling, pea-sized berriesand berries touching (BBCH 68, 75 and 79). Results are produced with leave-one-out cross validations where models are estimated from only 15 images per stage containing about 3.2×106 labellised pixels. The resulting classification performances are measured in terms of recall and precision that reached overall between 85% and 95% depending on the stage while overall accuracies range between 88% and 93%.

Estimation of terrain’s linear deformation rates using synthetic aperture radar systems

Synthetic Aperture Radars (SAR) are currently one of the most popular systems in the remote sensing domain, being widely utilized in the earth observation field. Their range of applicability extends in both marine and terrestrial regions. In the maritime domain, SAR systems are intensively used for the study of oceanic waves, waves breaking, marine currents, underwater topography, oil stains, for monitoring the glacier’s ice flow, and also for ships detection and localization. In the land areas, a class of applications which exploits the coherence propriety of SAR signals is able to retrieve information related to terrain’s characteristics, like topography and displacements. In this work, a processing chain for linear deformation rates estimation is presented and implemented on a dataset of 30 SAR images of Buzău and Focşani cities regions (Romania). The algorithm is based on identification of targets with stable electromagnetic response, exploiting their temporal coherence to obtain reliable estimates. An iterative phase regression analysis is conducted exclusively in the set of detected stable targets. The main challenge is represented by the estimation of the residual component of the phase, due to its random nature. Main feature of the proposed processing chain consists in the fact that it includes a step for terrain’s topography estimation, instead of using an external digital elevation model.

Azimuth Ambiguities Removal in Littoral Zones Based on Multi-Temporal SAR Images

Synthetic aperture radar (SAR) is one of the most important techniques for ocean monitoring. Azimuth ambiguities are a real problem in SAR images today, which can cause performance degradation in SAR ocean applications. In particular, littoral zones can be strongly affected by land-based sources, whereas they are usually regions of interest (ROI). Given the presence of complexity and diversity in littoral zones, azimuth ambiguities removal is a tough problem. As SAR sensors can have a repeat cycle, multi-temporal SAR images provide new insight into this problem. A method for azimuth ambiguities removal in littoral zones based on multi-temporal SAR images is proposed in this paper. The proposed processing chain includes co-registration, local correlation, binarization, masking, and restoration steps. It is designed to remove azimuth ambiguities caused by fixed land-based sources. The idea underlying the proposed method is that sea surface is dynamic, whereas azimuth ambiguities caused by land-based sources are constant. Thus, the temporal consistence of azimuth ambiguities is higher than sea clutter. It opens up the possibilities to use multi-temporal SAR data to remove azimuth ambiguities. The design of the method and the experimental procedure are based on images from the Sentinel data hub of Europe Space Agency (ESA). Both Interferometric Wide Swath (IW) and Stripmap (SM) mode images are taken into account to validate the proposed method. This paper also presents two RGB composition methods for better azimuth ambiguities visualization. Experimental results show that the proposed method can remove azimuth ambiguities in littoral zones effectively.

TOWARDS TIME-SERIES PROCESSING OF VHR SATELLITE IMAGES FOR SURFACE DEFORMATION DETECTON AND MEASUREMENTS

Abstract. The increasing fleet of VHR optical satellites (e.g. Pléiades, Spot 6/7, WorldView-3) offers new opportunities for the monitoring of surface deformation resulting from gravitational (e.g. glaciers, landslides) or tectonic forces (coseismic slip). Image correlation techniques have been developed and successfully employed in many geoscientific studies to quantify horizontal surface deformation at sub-pixel precision. The analysis of time-series, however, has received less attention in this context and there is still a lack of techniques that fully exploit archived image time-series and the increasing flux of incoming data. This study targets the development of an image correlation processing chain that relies on multiple pair-wise matching to exploit the redundancy of deformation measurements recorded at different view angles and over multiple time steps. The proposed processing chain is based on a hierarchical image correlation scheme that readily uses parallel processing. Since pair-wise matching can be performed independently the distribution of individual tasks is straightforward and yields to significant reductions of the overall runtime scaling with the available HPC infrastructure. We find that it is more convenient to implement experimental analytical tasks in a high-level programming language (i.e. R) and explore the use of parallel programming to compensate for performance bottlenecks of the interpreted language. Preliminary comparisons against maps from domain expert suggest that the proposed methodology is suitable to eliminate false detections and, thereby, enhances the reliability of correlation-based detections of surface deformation.

Discrimination of cell cycle phases in PCNA-immunolabeled cells.

BackgroundProtein function in eukaryotic cells is often controlled in a cell cycle-dependent manner. Therefore, the correct assignment of cellular phenotypes to cell cycle phases is a crucial task in cell biology research. Nuclear proteins whose localization varies during the cell cycle are valuable and frequently used markers of cell cycle progression. Proliferating cell nuclear antigen (PCNA) is a protein which is involved in DNA replication and has cell cycle dependent properties. In this work, we present a tool to identify cell cycle phases and in particular, sub-stages of the DNA replication phase (S-phase) based on the characteristic patterns of PCNA distribution. Single time point images of PCNA-immunolabeled cells are acquired using confocal and widefield fluorescence microscopy. In order to discriminate different cell cycle phases, an optimized processing pipeline is proposed. For this purpose, we provide an in-depth analysis and selection of appropriate features for classification, an in-depth evaluation of different classification algorithms, as well as a comparative analysis of classification performance achieved with confocal versus widefield microscopy images.ResultsWe show that the proposed processing chain is capable of automatically classifying cell cycle phases in PCNA-immunolabeled cells from single time point images, independently of the technique of image acquisition. Comparison of confocal and widefield images showed that for the proposed approach, the overall classification accuracy is slightly higher for confocal microscopy images.ConclusionOverall, automated identification of cell cycle phases and in particular, sub-stages of the DNA replication phase (S-phase) based on the characteristic patterns of PCNA distribution, is feasible for both confocal and widefield images.

Automated Underwater Object Recognition by Means of Fluorescence LIDAR

This paper focuses on automated recognition of underwater objects by means of light detection and ranging (LIDAR) systems. Differently from most works involved in underwater object recognition with LIDAR, where objects are recognized by their shape, here the interest is distinguishing objects on the basis of physical/chemical properties of object materials. To this aim, laser-induced fluorescence (LIF) spectroscopy is exploited, and an ad hoc signal processing chain is presented to effectively analyze the LIF spectra extracted at the detected object-range. Specifically, the goal is that of automatically recognizing the detected object with respect to a database (DB) of objects of interest, which have been previously spectrally characterized by means of laboratory fluorescence measurements. To this aim, suitable physics-based methodologies are proposed to compensate the signal for water-column effects. A decision-theory-based framework is developed to approach spectral recognition of the detected object with respect to the object DB. Experimental results from a laboratory test-bed show that the proposed processing chain is effective at automatically recognizing objects submerged in an artificial water column at different depths, based on a diverse DB of sample materials. The presented approach is shown to provide great potential for automated object recognition in marine and other water environments.

Spaceborne SAR data for global urban mapping at 30 m resolution using a robust urban extractor

With more than half of the world population now living in cities and 1.4 billion more people expected to move into cities by 2030, urban areas pose significant challenges on local, regional and global environment. Timely and accurate information on spatial distributions and temporal changes of urban areas are therefore needed to support sustainable development and environmental change research. The objective of this research is to evaluate spaceborne SAR data for improved global urban mapping using a robust processing chain, the KTH-Pavia Urban Extractor. The proposed processing chain includes urban extraction based on spatial indices and Grey Level Co-occurrence Matrix (GLCM) textures, an existing method and several improvements i.e., SAR data preprocessing, enhancement, and post-processing. ENVISAT Advanced Synthetic Aperture Radar (ASAR) C-VV data at 30m resolution were selected over 10 global cities and a rural area from six continents to demonstrate the robustness of the improved method. The results show that the KTH-Pavia Urban Extractor is effective in extracting urban areas and small towns from ENVISAT ASAR data and built-up areas can be mapped at 30m resolution with very good accuracy using only one or two SAR images. These findings indicate that operational global urban mapping is possible with spaceborne SAR data, especially with the launch of Sentinel-1 that provides SAR data with global coverage, operational reliability and quick data delivery.

A New Method for Cross-Normalization and Multitemporal Visualization of SAR Images for the Detection of Flooded Areas

Whenever multitemporal synthetic aperture radar (SAR) images are available, precise calibration and perfect spatial registration are required to obtain a useful image for displaying changes that have occurred. SAR calibration is a very complex and sensitive problem; some errors may persist after calibration that interfere with subsequent steps in the data fusion and visualization process. Because of the strong histogram asymmetry of SAR images, due to the well-known non-Gaussian model of radar backscattering, traditional image preprocessing procedures cannot be used here. A novel specific preprocessing phase, the so-called “cross-calibration/normalization,” is proposed to solve this problem. This, in turn, facilitates image enhancement and the numerical comparison of different image takes together with data fusion and visualization processes. The proposed processing chain includes filtering, histogram truncation, and equalization steps applied in an adaptive way to the images in question. The design of the method and the experimental procedure is based on images from the Italian Cosmo/Skymed mission. Both Stripmap and Spotlight images are taken into account to test the algorithms at different spatial resolutions. This paper also presents an example application: the generation of a single flood picture, the so-called “fast-ready flood map,” from multitemporal SAR images. The maps are very quickly and automatically generated without user interaction to support the authorities in providing first aid to a population. Toward this end, RGB composition is used: pre-flood and post-flood images are combined into a color image to better identify the flooded areas in comparison with permanent water and other classes.

A Robust Processing Chain for Face Recognition under Varying Illumination

Abstract In order to make face recognition more reliable under varying illumination, a robust processing chain is presented in this paper. Most of the illumination normalization methods treat all face images in the same way without considering the specific illumination condition of each probe image. For the nearly well-lit face images, they may be misclassified after illumination normalization. But they can be correctly classified without illumination normalization. To address this problem, the illumination quality index (IQI) of face image is proposed. According to the IQI of a probe face image, it can be detemvned whether the illumination normalization should be applied to it. In the proposed processing chain, the probe face image needing no illumination normalization will directly be used for recognition using normalized correlation. Otherwise a new illumination normalization approach, based on the Retinex theory and the total variation under L2 norm (TVL2) constraint model, is conducted on it. The pro...

Detection of structural changes of the human brain in longitudinally acquired MR images by deformation field morphometry: Methodological analysis, validation and application

The progression of neurodegenerative diseases as well as healthy aging is accompanied by structural changes of the brain. These changes are often only subtle when considered over time intervals of several months. Therefore morphometrical techniques for their detection in longitudinally acquired MR images must be highly sensitive, and they require a careful validation. In the present study, a novel processing chain for a longitudinal analysis based on deformation field morphometry is described. Procedures for its quantitative validation are also reported: Deformation fields were computed for the simulation of non-linear, local structural changes of human brains. Applying these deformation fields to “original” MR images yielded deformed MR images. The volume changes defined by the deformation fields represented the standard, against which the results of the longitudinal analysis of each pair of original and deformed MR image were compared. The proposed processing chain enabled to localize and to quantify simulated local atrophies near the cortex as well as in deep brain structures. An exemplary analysis of serial MR images of a patient suffering from an atypical Parkinson syndrome (cortico-basal degeneration, CBD) and healthy control subjects is presented, showing a characteristic pattern of volume changes in the brain of the patient which is strikingly different from the controls' patterns of changes.

A shape-based approach to change detection of lakes using time series remote sensing images

Shape analysis has not been considered in remote sensing as extensively as in other pattern recognition applications. However, shapes such as those of geometric patterns in agriculture and irregular boundaries of lakes can be extracted from the remotely sensed imagery even at relatively coarse spatial resolutions. This paper presents a procedure for efficiently retrieving and representing shapes of interesting features in remotely sensed imagery using supervised classification, object recognition, parametric contour tracing, and proposed piecewise linear polygonal approximation techniques. In addition, shape similarity can be measured by means of a computationally efficient metric. The study was conducted on a time series of radiometric and geometric rectified Landsat Multispectral Scanner (MSS) images and Thematic Mapper (TM) images, covering the scenes containing lakes in the Nebraska Sand Hills region. The results validate the effectiveness of the proposed processing chain in change detection of lake shapes and show that shape similarity is an important parameter in quantitatively measuring the spatial variations of objects.