Search

Pedestrian positioning is still a challenging field of interest when not receiving any GNSS or other reference signals as it might be the case in indoor environments or tunnels. For professional applications, where it is not possible to rely on any infrastructure, a common technique is to mount Inertial Measurement Units (IMUs) on the foot or other parts of the body for positioning. IMU based techniques still suffer from the remaining drift especially when the environment is unknown and not re-visited, or when the pedestrian walks randomly in large areas. In order to overcome this problem the so called FootSLAM algorithm applied in this paper is extended to handle known locations. FootSLAM is a SLAM (Simultaneous Localization and Mapping) algorithm estimating the map of the environment while walking wearing an inertial sensor on the foot or at other locations of the body. With only few known locations the estimated FootSLAM map can be refined or corrected in the case of non-convergence in critical areas. Beside the derivation of the algorithm handling known locations in FootSLAM, the influence of different kinds of prior information on the FootSLAM algorithm is analyzed in this paper in terms of error and map quality performance.

International audience; Many computer vision problems such as recognition, image retrieval, and tracking require matching two images. Currently, ones try to find as reliable as possible matching techniques with a very little constraint of computational time. In this paper, we are interested in applying image matching technique into robotic problems such as Augmented Reality, Visual Servoing in which the computational time is a critical element. We propose in this paper a real time keypoint based matching method. The novelties of this method include a fast corner detector, a compact corner descriptor based on Principal Component Analysis (PCA) technique and an efficient matching with help of Approximate Nearest Neighbor (ANN) technique. We show that the method gives a very satisfying result on accuracy as well as the computational time. The matching algorithm is applied to command a robot in a visual servoing application. It works at 10-14Hz and is well robust to variations in 3D viewpoint and illumination.

The fusion of complementary information from co-registered multi-modal image data enables a more detailed and more robust understanding of an image scene or specific objects, and is important for several applications in the field of remote sensing. In this paper, the benefits of combining RGB, near infrared (NIR) and thermal infrared (TIR) aerial images for the task of semantic vehicle segmentation through deep neural networks are investigated. Therefore, RGB, NIR and TIR image triplets acquired by the Modular Aerial Camera System (MACS) are precisely co-registered through the application of a virtual camera system and subsequently used for the training of different neural network architectures. Various experiments were conducted to investigate the influence of the different sensor characteristics and an early or late fusion within the network on the quality of the segmentation results.

The automatic matching of corresponding regions in remote sensing imagery acquired by synthetic aperture radar (SAR) and optical sensors is a crucial pre-requesite for many data fusion endeavours such as target recognition, image registration, or 3D-reconstruction by stereogrammetry. Driven by the success of deep learning in conventional optical image matching, we have carried out extensive research with regard to deep matching for SAR-optical multi-sensor image pairs in the recent past. In this paper, we summarize the achieved findings, including different concepts based on (pseudo-)siamese convolutional neural network architectures, hard negative mining, alternative formulations of the underlying loss function, and creation of artificial images by generative adversarial networks. Based on data from state-of-the-art remote sensing missions such as TerraSAR-X, Prism, Worldview-2, and Sentinel-1/2, we show what is already possible today, while highlighting challenges to be tackled by future research endeavors.

Building semantic segmentation is a crucial task for building information modeling (BIM). Current research generally exploits terrestrial image data, which provides only limited view of a building. By contrast, oblique imagery acquired by unmanned aerial vehicle (UAV) can provide richer information of both the building and its surroundings at a larger scale. In this paper, we present a novel pipeline for building semantic segmentation from oblique UAV images using a fully convolutional neural network (FCN). To cope with the lack of UAV image annotations at facade level, we leverage existing ground-view facades databases to simulate various aerial-view images based on estimated homography, yielding abundant synthetic aerial image annotations as training data. The FCN is trained end-to-end and tested on full-tile UAV images. Experiments demonstrate that the incorporation of simulated views can significantly boost the prediction accuracy of the network on UAV images and achieve reasonable segmentation performance.

This paper presents a system calibration method for a trifocal sensor, which is sensitive to different spectral bands. The trifocal camera system consists of a stereo camera, operating in the visual (VIS) spectrum and a thermal imaging camera, operating in the Long-Wave- Infrared (LWIR) spectrum. Intrinsic parameters and spatial alignment are determined simultaneously. As calibration target a passive aluminium chessboard is used. Corner detection and subsequent bundle adjustment is done on all synchronized image triplets. The remaining reprojection errors are in the sub-pixel range and enable the system to generate metric point clouds, colored with thermal intensities in real-time.

In this paper, we introduce the problem of cross-modal visuo-tactile object recognition with robotic active exploration. With this term, we mean that the robot observes a set of objects with visual perception, and later on, it is able to recognize such objects only with tactile exploration, without having touched any object before. Using a machine learning terminology, in our application, we have a visual training set and a tactile test set, or vice versa. To tackle this problem, we propose an approach constituted by four steps: finding a visuo-tactile common representation, defining a suitable set of features, transferring the features across the domains, and classifying the objects. We show the results of our approach using a set of 15 objects, collecting 40 visual examples and five tactile examples for each object. The proposed approach achieves an accuracy of 94.7%, which is comparable with the accuracy of the monomodal case, i.e., when using visual data both as training set and test set. Moreover, it performs well compared to the human ability, which we have roughly estimated carrying out an experiment with ten participants.

Summary Inverted pendulum on cart belongs to a class of highly non-linear, multivariable and complex system which is currently a keen area of interest for researchers in the field of control engineering and robotics. This paper considers stabilisation and control of inverted pendulum on cart moving on an inclined surface. The task of controlling inverted pendulum on inclined surface is much difficult as compared to that moving on horizontal surface. An offline control of the proposed system has been successfully achieved using proportional-integral-derivative (PID) and fuzzy controllers. A Matlab-Simulink model of the proposed system has been built using mathematical equations derived from Newton's second law. The study compares the proposed techniques in terms of settling time, overshoot, undershoot and steady state error. The simulation results are illustrated with the help of graphs and tables which confirms the validity of the proposed techniques.

Summary Breast cancer is significant health problem diagnosed mostly in women worldwide. Therefore, early detection of breast cancer is performed with the help of digital mammography, which can reduce mortality rate. This paper presents wrapper based feature selection approach for wavelet co-occurrence feature (WCF) using Genetic Fuzzy System (GFS) in mammogram classification problem. The performance of GFS algorithm is explained using mini-MIAS database. WCF features are obtained from detail wavelet coefficients at each level of decomposition of mammogram image. At first level of decomposition, 18 features are applied to GFS algorithm, which selects 5 features with an average classification success rate of 39.64%. Subsequently, at second level it selects 9 features from 36 features and the classification success rate is improved to 56.75%. For third level, 16 features are selected from 54 features and average success rate is improved to 64.98%. Lastly, at fourth level 72 features are applied to GFS, which selects 16 features and thereby increasing average success rate to 89.47%. Hence, GFS algorithm is the effective way of obtaining optimal set of feature in breast cancer diagnosis.

Summary This paper presents a technique that aimed to accomplish an efficient balance between video compression using H.265 protocol and retention of 8K resolution. The study implements multi-level of optimization in the encoding process using H.265 where JPEG2000 standards play a crucial role. The study also applies a novel concept of orthogonal projection that manages pixels metadata required in every frame transition followed by motion compensation. By using multiple file formats of 30 video datasets, the outcome of the study is found to be accomplishing approximately 49% of enhancement in data quality and around 59% of improvement in video compression in comparison to the existing techniques of HEVC-based video compression.

Summary In presence of environmental noise, speakers tend to emphasize their vocal effort to improve the audibility of voice. This involuntary adjustment is known as Lombard effect (LE). Due to LE the signal to noise ratio of speech increases, but at the same time the loudness, pitch and duration of phonemes changes. Hence, accuracy of automatic speech recognition systems degrades. In this paper, the effect of unsupervised equalization of Lombard effect is investigated for Hindi vowel classification task using Hindi database designed at TIFR Mumbai, India. Proposed Quantile-based Dynamic Cepstral Normalization MFCC (QCN-MFCC) along with baseline MFCC features have been used for vowel classification. Hidden Markov Model (HMM) is used as classifier. It is observed that QCN-MFCC features have given a maximum improvement of 5.97% and 5% over MFCC features for context-dependent and context-independent cases respectively. It is also observed that QCN-MFCC features have given improvement of 13% and 11.5% over MFCC features for context-dependent and context-independent classification of mid vowels.

Modern spaceborne synthetic aperture radar (SAR) sensors, such as TerraSAR-X/TanDEM-X and COSMO-SkyMed, can deliver very high resolution (VHR) data beyond the inherent spatial scales (on the order of 1m) of buildings, constituting invaluable data source for large-scale urban mapping. Processing this VHR data with advanced interferometric techniques, such as SAR tomography (TomoSAR), enables the generation of 3-D (or even 4-D) TomoSAR point clouds from space. In this paper, we present a novel and generic workflow that exploits these TomoSAR point clouds in a way that is capable to automatically produce benchmark annotated (buildings/non-buildings) SAR datasets. These annotated datasets (building masks) have been utilized to construct and train the state-of-the-art deep Fully Convolution Neural Networks with an additional Conditional Random Field represented as a Recurrent Neural Network to detect building regions in a single VHR SAR image. The results of building detection are illustrated and validated over TerraSAR-X VHR spotlight SAR image covering approximately 39 km2 - almost the whole city of Berlin - with mean pixel accuracies of around 93.84%.

Significantly outperforming traditional machine learning methods, deep convolutional neural networks have gained increasing popularity in the application of image classification and segmentation. Nevertheless, deep learning-based methods usually require a large amount of training data, which is quite labor-intensive and time-demanding. To deal with the problem in generating training data, we propose in this paper a novel approach to generate image annotations by transferring labels from aerial images to UAV images and refine the annotations using a densely connected CRF model with an embedded naive Bayes classifier. The generated annotations not only present correct semantic labels, but also preserve accurate class boundaries. To validate the utility of these automatic annotations, we deploy them as training data for pixel-wise image segmentation and compare the results with the segmentation using manual annotations. Experiment results demonstrate that the automatic annotations can achieve comparable segmentation accuracy as the manual annotations.

Exploratory visual analysis is often required to assist human operator to understand and interpret Earth Observation (EO) images. Optimal image representation offers cognitive support in discovering relevant facts about the scene with respect to a particular application. This is of crucial importance for training data sets selection in all Machine Learning tasks, particularly in the design of active learning tools for multispectral (MS) EO data. This paper proposes a deep neural network (DNN) based method to compress, learn and reveal the most significant information included in the spectral bands of EO data in support of relevant visualization for image content analysis. The advanced method uses a DNN to discover the most suggestive pseudo-color representation able to highlight the entire MS image content better than the particular 3 bands selection (R, G, B). We propose the use of information theory and the concept of mutual information to rank the spectral bands based on the amount of information contained, by applying the minimum-redundancy-maximum-relevance (mRMR) criterion on a the image so that we obtain the ranked bands. A DNN stacked autoencoder based paradigm is developed in order to extract and compress in three bands the overall information from the MS EO data. The developed method is demonstrated and validated for Sentinel 2 dataset.

Ultra-wideband (UWB) is a promising technology for positioning systems that has undergone massive research development in recent years. Most UWB positioning systems assume prior knowledge on the positions of the UWB anchors. Without knowing the UWB anchor positions, an accurate position estimate of an UWB tag is difficult. Hence, this paper presents a novel cooperative pedestrian simultaneous localization and mapping (SLAM) algorithm using an UWB positioning system without the necessity to have the prior information on the UWB anchor positions. To obtain accurate estimates of the UWB anchor positions, the estimations of the UWB anchor positions are shared between different mobile units. An increased accuracy of the estimated UWB anchor position relates directly to a more accurate position estimate of the mobile unit position. We evaluate the proposed algorithm based on measurements with multiple mobile units and fixed anchors with unknown positions. The evaluations show that an accurate position estimation of both the mobile units and the anchors is possible without any prior knowledge on the UWB anchor positions.

Extinction profile (EP) is an effective feature extraction method which can well preserve the geometrical characteristics of a hyperspectral image (HSI) and by extracting the EP from first three independent components (ICs) of an HSI, three correlated and complementary groups of EP features can be constructed. In this paper, an EPs fusion (EPs-F) strategy is proposed for HSI classification by exploring spatial-spectral information within and among three EP features. In general, the EPs-F method includes two stages. In the first stage, within each EP feature, a superpixel-based composite kernel strategy is proposed to adaptively fuse the spatial information of EP and the spectral feature of HSI. Then, the obtained adaptive composite kernel is used to create a classification map for each EP. In the second stage, decision fusion is further applied on different classification maps to create the final classification result. Experiments on two real HSIs verify the effectiveness of the proposed EPs-F algorithm.

This paper compares different similarity measures for the matching of very-high-resolution SAR and optical images over urban areas. It is meant to provide guidance about the performance of both signal-based and descriptor-based similarity measures in the context of this non-trivial case of multi-sensor correspondence matching. Using an automatically generated training dataset, thresholds for the distinction between correct matches and wrong matches are determined. It is shown that descriptor-based similarity measures outperform signal-based similarity measures significantly.

A major research area in remote sensing is the problem of multi-sensor data fusion. Especially the combination of images acquired by different sensor types, e.g. active and passive, is a difficult task. Over the last years deep learning methods have proven their high potential for remote sensing applications. In this paper we will show how a deep learning method can be valuable for the problem of optical and SAR image matching. We investigate the possible of conditional generative adversarial networks (cGANs) for the generation of artificial templates. Contrary to common template generation approaches for image matching, the generation of templates using cGANs does not require the extraction of features. Our results show the possibility of realistic SAR-like template generation from optical images through cGANs and the potential of these templates for enhancing the matching of optical and SAR images by means of reliability and accuracy.

The side-looking imaging geometry of synthetic aperture radar (SAR) causes inevitable layover in SAR images. Separating the contributions from different scatterers has been the fundamental for many applications. It is typically solved by explicit inversion of the SAR imaging model to retrieve the scattering profile along the mixed dimension (elevation), which is otherwise known as SAR tomography. This paper proposed a robust blind scatterer separation method to demix the layovered scatterers, avoiding the computationally expensive tomographic inversion. We demonstrate that the state-of-the-art principle component decomposition-based methods are heavily influenced by the nonergodicity of the selected samples, especially in urban area, such as point scatterers appearing often on facades. The proposed method is shown to be more robust than the state-of-the-art. Real data example shows that the proposed method outperforms the state-of-the-art by a factor of three in terms of the accuracy of the retrieved phase.

In this paper, we present a method for regularizing noisy 3D reconstructions, which is especially well suited for scenes containing planar structures like buildings. At horizontal structures, the input model is divided into slices and for each slice, an inside/outside labeling is computed. With the outlines of each slice labeling, we create an irregularly shaped volumetric cell decomposition of the whole scene. Then, an optimized inside/outside labeling of these cells is computed by solving an energy minimization problem. For the cell labeling optimization we introduce a novel smoothness term, where lines in the images are used to improve the regularization result. We show that our approach can take arbitrary dense meshed point clouds as input and delivers well regularized building models, which can be textured afterwards.

Deformation monitoring by multi-baseline synthetic aperture radar (SAR) interferometry is so far the only imaging-based method to assess millimeter-level deformation over large areas from space. Past research mostly focused on optimal deformation parameters retrieval on a pixel-basis. Only until recently, the first demonstration of object-based urban infrastructures monitoring by fusing SAR interferometry (InSAR) and the semantic classification labels derived from optical images was presented in [1]–[3]. This paper proposes an algorithm for object-based joint InSAR deformation reconstruction using these classification labels. We derive an object-based multi-baseline InSAR reconstruction model, and propose an efficient algorithm for bridge detection in optical images.

GNSS/IMU navigation systems offer low-cost and robust solution to navigate UAVs. Since redundant measurements greatly improve the reliability of navigation systems, extensive researches have been made to enhance the efficiency and robustness of GNSS/IMU by additional sensors. This paper presents a method for integrating reference data, images taken from UAVs, barometric height data and GNSS/IMU data to estimate accurate and reliable pose parameters of UAVs. We provide improved pose estimations by integrating multi-sensor observations in an EKF algorithm with IMU motion model. The implemented methodology has demonstrated to be very efficient and reliable for automatic pose estimation. The calculated position and attitude of the UAV especially when we removed the GNSS from the working cycle clearly indicate the ability of the purposed methodology.

Fault localization in RTL design using dynamic dependency graphs.Reverse debugging for RTL designs.Scalable and fast approach applicable to large designs.Manual debugging effort can be reduced by more than 50ź%. Display Omitted Debugging is a time consuming task in hardware design. In this paper a new debugging approach based on the analysis of dynamic dependency graphs is presented. Powerful techniques for software debugging, including reverse debugging, dynamic forward and backward slicing, and spectrum-based fault localization are combined and adapted for hardware designs. A case study on designs with multiple faults approved the power of the proposed debugging methodology reducing the debugging time to 50% in comparison to conventional techniques.

This paper presents monostatic radar cross section (RCS) measurements of a human hand in the frequency range from 60GHz to 90 GHz. These values are important parameters for system designs in the emerging field of radar-based gesture recognition. The measurement procedure is described in detail and results of four gestures at three different distances are depicted and analysed., acceptedVersion

In this paper, we propose a new scheme for script and nature identification. The objective is to discriminate between machine-printed/handwritten and Latin/Arabic scripts at word level. It is relatively a complex task due to possible use of multi-fonts and sizes, complexity and variation in handwriting. In the proposed script identification system, we extract features from word images using Co-occurrence Matrix of Oriented Gradients (Co-MOG). The classification is done using different classifiers. Extensive experimentation has been carried on 24000 words, extracted from standard databases. An average identification accuracy of 99.85% is achieved by k Nearest Neighbors (k-NN) classifier which clearly outperforms results of some existing systems.

Very high resolution satellite imagery used to be a rare commodity, with infrequent satellite pass-over times over a specific area-of-interest obviating many useful applications. Today, more and more such satellite systems are available, with visual analysis and interpretation of imagery still important to derive relevant features and changes from satellite data. In order to allow efficient, robust and routine image analysis for humanitarian purposes, semi-automated feature extraction is of increasing importance for operational emergency mapping tasks. In the frame of the European Earth Observation program COPERNICUS and related research activities under the European Union's Seventh Framework Program, substantial scientific developments and mapping services are dedicated to satellite based humanitarian mapping and monitoring. In this paper, recent results in methodological research and development of routine services in satellite mapping for humanitarian situational awareness are reviewed and discussed. Ethical aspects of sensitivity and security of humanitarian mapping are deliberated. Furthermore methods for monitoring and analysis of refugee/internally displaced persons camps in humanitarian settings are assessed. Advantages and limitations of object-based image analysis, sample supervised segmentation and feature extraction are presented and discussed.

This paper addresses the imaging of objects located under a forest cover using polarimetric synthetic aperture radar tomography (POLTOMSAR) at L-band. High-resolution spectral estimators, able to accurately discriminate multiple scattering centers in the vertical direction, are used to separate the response of objects and vehicles embedded in a volumetric background. A new polarimetric spectral analysis technique is introduced and is shown to improve the estimation accuracy of the vertical position of both artificial scatterers and natural environments. This approach provides optimal polarimetric features that may be used to further characterize the objects under analysis. The effectiveness of this novel technique for POLTOMSAR is demonstrated using fully polarimetric L-band airborne data sets acquired by the German Aerospace Center (DLR)'s E-SAR system over the test site in Dornstetten, Germany.

This paper shows how to reduce evaluation time for context inference. Probabilistic Context Inference has proven to be a good representation of the physical reality with uncertain or missing information, giving with the probability also a measure of the quality of information. As the inference complexity is very high, the complexity of the to be evaluated rule (representing a share of the real world) should be reduced as far as possible. Therefore we present an approach to select only relevant values of context types and to adapt this selection during its usage time. A short proof of concept indicates that both targets, reducing inference time and maintaining quality of information, can be reached with the proposed approach.

Unsupervised classification is an essential step in the automatic analysis of SAR remote sensing data. Classification results make SAR data easier to interpret and can serve as a starting point for automated analysis techniques that apply to homogeneous regions of the observed scene. Polarimetric SAR data are particularly interesting for unsupervised classification purposes, since they contain a great amount of information, allowing robust statistical clustering of the image content on the one hand and a direct physical interpretation of the result on the other. This paper proposes a new unsupervised classification approach for polarimetric SAR data. Assuming Wishart-distributed polarimetric covariance matrices, it combines spectral clustering based on the covariance matrices themselves with spatial clustering by statistical analysis of local neighbourhoods. Instead of working with binary assignments of samples to class centres, a soft decision rule is used in which each pixel is assigned to all class centres in the spectral and spatial domains. The local neighbourhood is taken into account by altering the probabilities of class membership by a neighbourhood function, obtained from normalized compatibility coefficients, describing cluster sizes and mutual tolerance. In this way, robust and homogenous classification results can be obtained even in the presence of strong speckle noise.

Artificial intelligence and expert systems are the latest buzzwords and the hottest topics in the scientific community today Some experts are proclaiming that artificial intelligence (AI) has already emerged as one of the most significant technologies of this cen tury. Proponents are declaring that it will completely revolu tionize management and the way computers are used. If these claims are even half true, then AI is bound to have a profound effect upon the art and science of simulation. This paper will provide an overview of this rapidly evolving field, examine the potential ofAl (and more particularly, expert systems) in simula tion and attempt to explore the probable impact as well as fore cast likely future directions.

In this paper, we present the first conditional generative adversarial network (cGAN) architecture that is specifically designed to fuse synthetic aperture radar (SAR) and optical multi-spectral (MS) image data to generate cloud- and haze-free MS optical data from a cloud-corrupted MS input and an auxiliary SAR image. Experiments on Sentinel-2 MS and Sentinel-l SAR data confirm that our extended SAR-Opt-cGAN model utilizes the auxiliary SAR information to better reconstruct MS images than an equivalent model which uses the same architecture but only single-sensor MS data as input.