Sistema de detecção e localização de veículos subaquáticos

Mestrado em Engenharia Electrotécnica e de Computadores.Área de Especialização de Sistemas Autónomos

Sistema visual para Ground Truth de sistemas autónomos

Este trabalho visa contribuir para o desenvolvimento de um sistema de visão multi-câmara para determinação da localização, atitude e seguimento de múltiplos objectos, para ser utilizado na unidade de robótica do INESCTEC, e resulta da necessidade de ter informação externa exacta que sirva de referência no estudo, caracterização e desenvolvimento de algoritmos de localização, navegação e controlo de vários sistemas autónomos. Com base na caracterização dos veículos autónomos existentes na unidade de robótica do INESCTEC e na análise dos seus cenários de operação, foi efectuado o levantamento de requisitos para o sistema a desenvolver. Foram estudados os fundamentos teóricos, necessários ao desenvolvimento do sistema, em temas relacionados com visão computacional, métodos de estimação e associação de dados para problemas de seguimento de múltiplos objectos . Foi proposta uma arquitectura para o sistema global que endereça os vários requisitos identi cados, permitindo a utilização de múltiplas câmaras e suportando o seguimento de múltiplos objectos, com ou sem marcadores. Foram implementados e validados componentes da arquitectura proposta e integrados num sistema para validação, focando na localização e seguimento de múltiplos objectos com marcadores luminosos à base de Light-Emitting Diodes (LEDs). Nomeadamente, os módulos para a identi cação dos pontos de interesse na imagem, técnicas para agrupar os vários pontos de interesse de cada objecto e efectuar a correspondência das medidas obtidas pelas várias câmaras, método para a determinação da posição e atitude dos objectos, ltro para seguimento de múltiplos objectos. Foram realizados testes para validação e a nação do sistema implementado que demonstram que a solução encontrada vai de encontro aos requisitos, e foram identi cadas as linhas de trabalho para a continuação do desenvolvimento do sistema global.

Groundtruth system for underwater benchmarking

Oceans - San Diego, 2013

Mechanical characterization of aortas using 2D ultrasound elastography

Rupture of aortic aneurysms (AA) is a major cause of death in the Western world. Currently, clinical decision upon surgical intervention is based on the diameter of the aneurysm. However, this method is not fully adequate. Noninvasive assessment of the elastic properties of the arterial wall can be a better predictor for AA growth and rupture risk. The purpose of this study is to estimate mechanical properties of the aortic wall using in vitro inflation testing and 2D ultrasound (US) elastography, and investigate the performance of the proposed methodology for physiological conditions. Two different inflation experiments were performed on twelve porcine aortas: 1) a static experiment for a large pressure range (0 – 140 mmHg); 2) a dynamic experiment closely mimicking the in vivo hemodynamics at physiological pressures (70 – 130 mmHg). 2D raw radiofrequency (RF) US datasets were acquired for one longitudinal and two cross-sectional imaging planes, for both experiments. The RF-data were manually segmented and a 2D vessel wall displacement tracking algorithm was applied to obtain the aortic diameter–time behavior. The shear modulus G was estimated assuming a Neo-Hookean material model. In addition, an incremental study based on the static data was performed to: 1) investigate the changes in G for increasing mean arterial pressure (MAP), for a certain pressure difference (30, 40, 50 and 60 mmHg); 2) compare the results with those from the dynamic experiment, for the same pressure range. The resulting shear modulus G was 94 ± 16 kPa for the static experiment, which is in agreement with literature. A linear dependency on MAP was found for G, yet the effect of the pressure difference was negligible. The dynamic data revealed a G of 250 ± 20 kPa. For the same pressure range, the incremental shear modulus (Ginc) was 240 ± 39 kPa, which is in agreement with the former. In general, for all experiments, no significant differences in the values of G were found between different image planes. This study shows that 2D US elastography of aortas during inflation testing is feasible under controlled and physiological circumstances. In future studies, the in vivo, dynamic experiment should be repeated for a range of MAPs and pathological vessels should be examined. Furthermore, the use of more complex material models needs to be considered to describe the non-linear behavior of the vascular tissue.

Validation of a clinical algorithm to identify low-risk patients with pulmonary embolism.

Summary Background: We previously derived a clinical prognostic algorithm to identify patients with pulmonary embolism (PE) who are at low-risk of short-term mortality who could be safely discharged early or treated entirely in an outpatient setting. Objectives: To externally validate the clinical prognostic algorithm in an independent patient sample. Methods: We validated the algorithm in 983 consecutive patients prospectively diagnosed with PE at an emergency department of a university hospital. Patients with none of the algorithm's 10 prognostic variables (age >/= 70 years, cancer, heart failure, chronic lung disease, chronic renal disease, cerebrovascular disease, pulse >/= 110/min., systolic blood pressure < 100 mm Hg, oxygen saturation < 90%, and altered mental status) at baseline were defined as low-risk. We compared 30-day overall mortality among low-risk patients based on the algorithm between the validation and the original derivation sample. We also assessed the rate of PE-related and bleeding-related mortality among low-risk patients. Results: Overall, the algorithm classified 16.3% of patients with PE as low-risk. Mortality at 30 days was 1.9% among low-risk patients and did not differ between the validation and the original derivation sample. Among low-risk patients, only 0.6% died from definite or possible PE, and 0% died from bleeding. Conclusions: This study validates an easy-to-use, clinical prognostic algorithm for PE that accurately identifies patients with PE who are at low-risk of short-term mortality. Low-risk patients based on our algorithm are potential candidates for less costly outpatient treatment.

Oscillatory magnetic tweezers based on ferromagnetic beads and simple coaxial coils

We report the design and validation of simple magnetic tweezers for oscillating ferromagnetic beads in the piconewton and nanometer scales. The system is based on a single pair of coaxial coils operating in two sequential modes: permanent magnetization of the beads through a large and brief pulse of magnetic field and generation of magnetic gradients to produce uniaxial oscillatory forces. By using this two step method, the magnetic moment of the beads remains constant during measurements. Therefore, the applied force can be computed and varies linearly with the driving signal. No feedback control is required to produce well defined force oscillations over a wide bandwidth. The design of the coils was optimized to obtain high magnetic fields (280 mT) and gradients (2 T/m) with high homogeneity (5% variation) within the sample. The magnetic tweezers were implemented in an inverted optical microscope with a videomicroscopy-based multiparticle tracking system. The apparatus was validated with 4.5 ¿m magnetite beads obtaining forces up to ~2 pN and subnanometer resolution. The applicability of the device includes microrheology of biopolymer and cell cytoplasm, molecular mechanics, and mechanotransduction in living cells.

Cross-frequency coupling in EEG revealed by adaptive oscillation tracking.

Introduction: Neuronal oscillations have been the focus of increasing interest in the neuroscientific community, in part because they have been considered as a possible integrating mechanism through which internal states can influence stimulus processing in a top-down way (Engel et al., 2001). Moreover, increasing evidence indicates that oscillations in different frequency bands interact with one other through coupling mechanisms (Jensen and Colgin, 2007). The existence and the importance of these cross-frequency couplings during various tasks have been verified by recent studies (Canolty et al., 2006; Lakatos et al., 2007). In this study, we measure the strength and directionality of two types of couplings - phase-amplitude couplings and phase-phase couplings - between various bands in EEG data recorded during an illusory contour experiment that were identified using a recently-proposed adaptive frequency tracking algorithm (Van Zaen et al., 2010). Methods: The data used in this study have been taken from a previously published study examining the spatiotemporal mechanisms of illusory contour processing (Murray et al., 2002). The EEG in the present study were from a subset of nine subjects. Each stimulus was composed of 'pac-man' inducers presented in two orientations: IC, when an illusory contour was present, and NC, when no contour could be detected. The signals recorded by the electrodes P2, P4, P6, PO4 and PO6 were averaged, and filtered into the following bands: 4-8Hz, 8-12Hz, 15-25Hz, 35-45Hz, 45-55Hz, 55-65Hz and 65-75Hz. An adaptive frequency tracking algorithm (Van Zaen et al., 2010) was then applied in each band in order to extract the main oscillation and estimate its frequency. This additional step ensures that clean phase information is obtained when taking the Hilbert transform. The frequency estimated by the tracker was averaged over sliding windows and then used to compare the two conditions. Two types of cross-frequency couplings were considered: phase-amplitude couplings and phase-phase couplings. Both types were measured with the phase locking value (PLV, Lachaux et al., 1999) over sliding windows. The phase-amplitude couplings were computed with the phase of the low frequency oscillation and the phase of the amplitude of the high frequency one. Different coupling coefficients were used when measuring phase-phase couplings in order to estimate different m:n synchronizations (4:3, 3:2, 2:1, 3:1, 4:1, 5:1, 6:1, 7:1, 8:1 and 9:1) and to take into account the frequency differences across bands. Moreover, the direction of coupling was estimated with a directionality index (Bahraminasab et al., 2008). Finally, the two conditions IC and NC were compared with ANOVAs with 'subject' as a random effect and 'condition' as a fixed effect. Before computing the statistical tests, the PLV values were transformed into approximately normal variables (Penny et al., 2008). Results: When comparing the mean estimated frequency across conditions, a significant difference was found only in the 4-8Hz band, such that the frequency within this band was significantly higher for IC than NC stimuli starting at ~250ms post-stimulus onset (Fig. 1; solid line shows IC and dashed line NC). Significant differences in phase-amplitude couplings were obtained only when the 4-8 Hz band was taken as the low frequency band. Moreover, in all significant situations, the coupling strength is higher for the NC than IC condition. An example of significant difference between conditions is shown in Fig. 2 for the phase-amplitude coupling between the 4-8Hz and 55-65Hz bands (p-value in top panel and mean PLV values in the bottom panel). A decrease in coupling strength was observed shortly after stimulus onset for both conditions and was greater for the condition IC. This phenomenon was observed with all other frequency bands. The results obtained for the phase-phase couplings were more complex. As for the phase-amplitude couplings, all significant differences were obtained when the 4-8Hz band was considered as the low frequency band. The stimulus condition exhibiting the higher coupling strength depended on the ratio of the coupling coefficients. When this ratio was small, the IC condition exhibited the higher phase-phase coupling strength. When this ratio was large, the NC condition exhibited the higher coupling strength. Fig. 3 shows the phase-phase couplings between the 4-8Hz and 35-45Hz bands for the coupling coefficient 6:1, and the coupling strength was significantly higher for the IC than NC condition. By contrast, for the coupling coefficient 9:1 the NC condition gave the higher coupling strength (Fig. 4). Control analyses verified that it is not a consequence of the frequency difference between the two conditions in the 4-8Hz band. The directionality measures indicated a transfer of information from the low frequency components towards the high frequency ones. Conclusions: Adaptive tracking is a feasible method for EEG analyses, revealing information both about stimulus-related differences and coupling patterns across frequencies. Theta oscillations play a central role in illusory shape processing and more generally in visual processing. The presence vs. absence of illusory shapes was paralleled by faster theta oscillations. Phase-amplitude couplings were decreased more for IC than NC and might be due to a resetting mechanism. The complex patterns in phase-phase coupling between theta and beta/gamma suggest that the contribution of these oscillations to visual binding and stimulus processing are not as straightforward as conventionally held. Causality analyses further suggest that theta oscillations drive beta/gamma oscillations (see also Schroeder and Lakatos, 2009). The present findings highlight the need for applying more sophisticated signal analyses in order to establish a fuller understanding of the functional role of neural oscillations.

Discriminating crop, weeds and soil surface with a terrestrial LIDAR sensor

In this study, the evaluation of the accuracy and performance of a light detection and ranging (LIDAR) sensor for vegetation using distance and reflection measurements aiming to detect and discriminate maize plants and weeds from soil surface was done. The study continues a previous work carried out in a maize field in Spain with a LIDAR sensor using exclusively one index, the height profile. The current system uses a combination of the two mentioned indexes. The experiment was carried out in a maize field at growth stage 12–14, at 16 different locations selected to represent the widest possible density of three weeds: Echinochloa crus-galli (L.) P.Beauv., Lamium purpureum L., Galium aparine L.and Veronica persica Poir.. A terrestrial LIDAR sensor was mounted on a tripod pointing to the inter-row area, with its horizontal axis and the field of view pointing vertically downwards to the ground, scanning a vertical plane with the potential presence of vegetation. Immediately after the LIDAR data acquisition (distances and reflection measurements), actual heights of plants were estimated using an appropriate methodology. For that purpose, digital images were taken of each sampled area. Data showed a high correlation between LIDAR measured height and actual plant heights (R2 = 0.75). Binary logistic regression between weed presence/absence and the sensor readings (LIDAR height and reflection values) was used to validate the accuracy of the sensor. This permitted the discrimination of vegetation from the ground with an accuracy of up to 95%. In addition, a Canonical Discrimination Analysis (CDA) was able to discriminate mostly between soil and vegetation and, to a far lesser extent, between crop and weeds. The studied methodology arises as a good system for weed detection, which in combination with other principles, such as vision-based technologies, could improve the efficiency and accuracy of herbicide spraying.

Sensor Fusion of Proprioception, Force and Vision in Estimation and Robot

Sensor-based robot control allows manipulation in dynamic environments with uncertainties. Vision is a versatile low-cost sensory modality, but low sample rate, high sensor delay and uncertain measurements limit its usability, especially in strongly dynamic environments. Force is a complementary sensory modality allowing accurate measurements of local object shape when a tooltip is in contact with the object. In multimodal sensor fusion, several sensors measuring different modalities are combined to give a more accurate estimate of the environment. As force and vision are fundamentally different sensory modalities not sharing a common representation, combining the information from these sensors is not straightforward. In this thesis, methods for fusing proprioception, force and vision together are proposed. Making assumptions of object shape and modeling the uncertainties of the sensors, the measurements can be fused together in an extended Kalman filter. The fusion of force and visual measurements makes it possible to estimate the pose of a moving target with an end-effector mounted moving camera at high rate and accuracy. The proposed approach takes the latency of the vision system into account explicitly, to provide high sample rate estimates. The estimates also allow a smooth transition from vision-based motion control to force control. The velocity of the end-effector can be controlled by estimating the distance to the target by vision and determining the velocity profile giving rapid approach and minimal force overshoot. Experiments with a 5-degree-of-freedom parallel hydraulic manipulator and a 6-degree-of-freedom serial manipulator show that integration of several sensor modalities can increase the accuracy of the measurements significantly.

Machine vision methods for process measurements in pulping

The papermaking industry has been continuously developing intelligent solutions to characterize the raw materials it uses, to control the manufacturing process in a robust way, and to guarantee the desired quality of the end product. Based on the much improved imaging techniques and image-based analysis methods, it has become possible to look inside the manufacturing pipeline and propose more effective alternatives to human expertise. This study is focused on the development of image analyses methods for the pulping process of papermaking. Pulping starts with wood disintegration and forming the fiber suspension that is subsequently bleached, mixed with additives and chemicals, and finally dried and shipped to the papermaking mills. At each stage of the process it is important to analyze the properties of the raw material to guarantee the product quality. In order to evaluate properties of fibers, the main component of the pulp suspension, a framework for fiber characterization based on microscopic images is proposed in this thesis as the first contribution. The framework allows computation of fiber length and curl index correlating well with the ground truth values. The bubble detection method, the second contribution, was developed in order to estimate the gas volume at the delignification stage of the pulping process based on high-resolution in-line imaging. The gas volume was estimated accurately and the solution enabled just-in-time process termination whereas the accurate estimation of bubble size categories still remained challenging. As the third contribution of the study, optical flow computation was studied and the methods were successfully applied to pulp flow velocity estimation based on double-exposed images. Finally, a framework for classifying dirt particles in dried pulp sheets, including the semisynthetic ground truth generation, feature selection, and performance comparison of the state-of-the-art classification techniques, was proposed as the fourth contribution. The framework was successfully tested on the semisynthetic and real-world pulp sheet images. These four contributions assist in developing an integrated factory-level vision-based process control.

Automatic traffic sign inventory- and condition analysis

This thesis researches automatic traffic sign inventory and condition analysis using machine vision and pattern recognition methods. Automatic traffic sign inventory and condition analysis can be used to more efficient road maintenance, improving the maintenance processes, and to enable intelligent driving systems. Automatic traffic sign detection and classification has been researched before from the viewpoint of self-driving vehicles, driver assistance systems, and the use of signs in mapping services. Machine vision based inventory of traffic signs consists of detection, classification, localization, and condition analysis of traffic signs. The produced machine vision system performance is estimated with three datasets, from which two of have been been collected for this thesis. Based on the experiments almost all traffic signs can be detected, classified, and located and their condition analysed. In future, the inventory system performance has to be verified in challenging conditions and the system has to be pilot tested.

Video-based analysis of Gait pathologies

L’analyse de la marche a émergé comme l’un des domaines médicaux le plus im- portants récemment. Les systèmes à base de marqueurs sont les méthodes les plus fa- vorisées par l’évaluation du mouvement humain et l’analyse de la marche, cependant, ces systèmes nécessitent des équipements et de l’expertise spécifiques et sont lourds, coûteux et difficiles à utiliser. De nombreuses approches récentes basées sur la vision par ordinateur ont été développées pour réduire le coût des systèmes de capture de mou- vement tout en assurant un résultat de haute précision. Dans cette thèse, nous présentons notre nouveau système d’analyse de la démarche à faible coût, qui est composé de deux caméras vidéo monoculaire placées sur le côté gauche et droit d’un tapis roulant. Chaque modèle 2D de la moitié du squelette humain est reconstruit à partir de chaque vue sur la base de la segmentation dynamique de la couleur, l’analyse de la marche est alors effectuée sur ces deux modèles. La validation avec l’état de l’art basée sur la vision du système de capture de mouvement (en utilisant le Microsoft Kinect) et la réalité du ter- rain (avec des marqueurs) a été faite pour démontrer la robustesse et l’efficacité de notre système. L’erreur moyenne de l’estimation du modèle de squelette humain par rapport à la réalité du terrain entre notre méthode vs Kinect est très prometteur: les joints des angles de cuisses (6,29◦ contre 9,68◦), jambes (7,68◦ contre 11,47◦), pieds (6,14◦ contre 13,63◦), la longueur de la foulée (6.14cm rapport de 13.63cm) sont meilleurs et plus stables que ceux de la Kinect, alors que le système peut maintenir une précision assez proche de la Kinect pour les bras (7,29◦ contre 6,12◦), les bras inférieurs (8,33◦ contre 8,04◦), et le torse (8,69◦contre 6,47◦). Basé sur le modèle de squelette obtenu par chaque méthode, nous avons réalisé une étude de symétrie sur différentes articulations (coude, genou et cheville) en utilisant chaque méthode sur trois sujets différents pour voir quelle méthode permet de distinguer plus efficacement la caractéristique symétrie / asymétrie de la marche. Dans notre test, notre système a un angle de genou au maximum de 8,97◦ et 13,86◦ pour des promenades normale et asymétrique respectivement, tandis que la Kinect a donné 10,58◦et 11,94◦. Par rapport à la réalité de terrain, 7,64◦et 14,34◦, notre système a montré une plus grande précision et pouvoir discriminant entre les deux cas.

Computer vision techniques in seafood quality control

The application of computer vision based quality control has been slowly but steadily gaining importance mainly due to its speed in achieving results and also greatly due to its non- destnictive nature of testing. Besides, in food applications it also does not contribute to contamination. However, computer vision applications in quality control needs the application of an appropriate software for image analysis. Eventhough computer vision based quality control has several advantages, its application has limitations as to the type of work to be done, particularly so in the food industries. Selective applications, however, can be highly advantageous and very accurate.Computer vision based image analysis could be used in morphometric measurements of fish with the same accuracy as the existing conventional method. The method is non-destructive and non-contaminating thus providing anadvantage in seafood processing.The images could be stored in archives and retrieved at anytime to carry out morphometric studies for biologists.Computer vision and subsequent image analysis could be used in measurements of various food products to assess uniformity of size. One product namely cutlet and product ingredients namely coating materials such as bread crumbs and rava were selected for the study. Computer vision based image analysis was used in the measurements of length, width and area of cutlets. Also the width of coating materials like bread crumbs was measured.Computer imaging and subsequent image analysis can be very effectively used in quality evaluations of product ingredients in food processing. Measurement of width of coating materials could establish uniformity of particles or the lack of it. The application of image analysis in bacteriological work was also done

ECG Noise Removal using GA tuned Sign-Data Least Mean Square Algorithm

Adaptive filter is a primary method to filter Electrocardiogram (ECG), because it does not need the signal statistical characteristics. In this paper, an adaptive filtering technique for denoising the ECG based on Genetic Algorithm (GA) tuned Sign-Data Least Mean Square (SD-LMS) algorithm is proposed. This technique minimizes the mean-squared error between the primary input, which is a noisy ECG, and a reference input which can be either noise that is correlated in some way with the noise in the primary input or a signal that is correlated only with ECG in the primary input. Noise is used as the reference signal in this work. The algorithm was applied to the records from the MIT -BIH Arrhythmia database for removing the baseline wander and 60Hz power line interference. The proposed algorithm gave an average signal to noise ratio improvement of 10.75 dB for baseline wander and 24.26 dB for power line interference which is better than the previous reported works

Internal Camera Calibration Using Rotation and Geometric Shapes

This paper describes a simple method for internal camera calibration for computer vision. This method is based on tracking image features through a sequence of images while the camera undergoes pure rotation. The location of the features relative to the camera or to each other need not be known and therefore this method can be used both for laboratory calibration and for self calibration in autonomous robots working in unstructured environments. A second method of calibration is also presented. This method uses simple geometric objects such as spheres and straight lines to The camera parameters. Calibration is performed using both methods and the results compared.