Fechamento de loops é um dos principais processos das estratégias de SLAM baseadas em grafos, usadas para estimar o erro de deslocamento acumulado à ser minimizado pela técnica. Neste sentido, boas correspondências de cenas permitem criar uma conexão entre dois nós do grafo que está sendo construído para representar o ambiente. Contudo, falsas correspondências podem levar essas estratégias a um estado irreversível de falsa representação do ambiente. Neste trabalho, um método robusto baseado em features que usa sequências de imagens para reconhecer áreas revisitadas é apresentado. Este método usa a abordagem de Bag-of-Words para reduzir efeitos de iluminação e uma ponderação TF-IDF para ressaltar as principais features que descrevem cada cena. Além disso, um algoritmo baseado na técnica de Mean Shift é usado sobre uma matriz de similaridade para identificar a possível trajetória seguida pelo robô e melhorar a detecção de fechamento de loop. O método apresentado foi testado em um ambiente aberto usando sequências de imagens coletadas com usando uma câmera de mão e um drone modelo Parrot ArDrone 2.0. / Loop closure recognition is one of the main processes of graph-based SLAM strategies, used to estimate the accumulated motion error to be minimized by the technique. Good scene correspondences allow to create constraints between two nodes in the graph that is currently being built to represent the environment that the robot is immersed. However, false correspondences can lead these strategies to an irreversible wrong environment representation. In this work, we present a robust feature-based loop closure approach that uses image sequence matching to recognize revisited areas. This approach uses Bag-of- Words to reduce the effects of lightning changes and a TF-IDF weighting to enhance the main features that describe each scene. Besides, an algorithm based on Mean Shift is used over a similarity matrix to identify the possible trajectory followed by the robot and improve the loop closure detection. Our method is tested in a GPS-denied outdoor environment using image sequences collected using a handheld camera and a Parrot ArDrone 2.0.

Fechamento de loops é um dos principais processos das estratégias de SLAM baseadas em grafos, usadas para estimar o erro de deslocamento acumulado à ser minimizado pela técnica. Neste sentido, boas correspondências de cenas permitem criar uma conexão entre dois nós do grafo que está sendo construído para representar o ambiente. Contudo, falsas correspondências podem levar essas estratégias a um estado irreversível de falsa representação do ambiente. Neste trabalho, um método robusto baseado em features que usa sequências de imagens para reconhecer áreas revisitadas é apresentado. Este método usa a abordagem de Bag-of-Words para reduzir efeitos de iluminação e uma ponderação TF-IDF para ressaltar as principais features que descrevem cada cena. Além disso, um algoritmo baseado na técnica de Mean Shift é usado sobre uma matriz de similaridade para identificar a possível trajetória seguida pelo robô e melhorar a detecção de fechamento de loop. O método apresentado foi testado em um ambiente aberto usando sequências de imagens coletadas com usando uma câmera de mão e um drone modelo Parrot ArDrone 2.0. / Loop closure recognition is one of the main processes of graph-based SLAM strategies, used to estimate the accumulated motion error to be minimized by the technique. Good scene correspondences allow to create constraints between two nodes in the graph that is currently being built to represent the environment that the robot is immersed. However, false correspondences can lead these strategies to an irreversible wrong environment representation. In this work, we present a robust feature-based loop closure approach that uses image sequence matching to recognize revisited areas. This approach uses Bag-of- Words to reduce the effects of lightning changes and a TF-IDF weighting to enhance the main features that describe each scene. Besides, an algorithm based on Mean Shift is used over a similarity matrix to identify the possible trajectory followed by the robot and improve the loop closure detection. Our method is tested in a GPS-denied outdoor environment using image sequences collected using a handheld camera and a Parrot ArDrone 2.0.

Fechamento de loops é um dos principais processos das estratégias de SLAM baseadas em grafos, usadas para estimar o erro de deslocamento acumulado à ser minimizado pela técnica. Neste sentido, boas correspondências de cenas permitem criar uma conexão entre dois nós do grafo que está sendo construído para representar o ambiente. Contudo, falsas correspondências podem levar essas estratégias a um estado irreversível de falsa representação do ambiente. Neste trabalho, um método robusto baseado em features que usa sequências de imagens para reconhecer áreas revisitadas é apresentado. Este método usa a abordagem de Bag-of-Words para reduzir efeitos de iluminação e uma ponderação TF-IDF para ressaltar as principais features que descrevem cada cena. Além disso, um algoritmo baseado na técnica de Mean Shift é usado sobre uma matriz de similaridade para identificar a possível trajetória seguida pelo robô e melhorar a detecção de fechamento de loop. O método apresentado foi testado em um ambiente aberto usando sequências de imagens coletadas com usando uma câmera de mão e um drone modelo Parrot ArDrone 2.0. / Loop closure recognition is one of the main processes of graph-based SLAM strategies, used to estimate the accumulated motion error to be minimized by the technique. Good scene correspondences allow to create constraints between two nodes in the graph that is currently being built to represent the environment that the robot is immersed. However, false correspondences can lead these strategies to an irreversible wrong environment representation. In this work, we present a robust feature-based loop closure approach that uses image sequence matching to recognize revisited areas. This approach uses Bag-of- Words to reduce the effects of lightning changes and a TF-IDF weighting to enhance the main features that describe each scene. Besides, an algorithm based on Mean Shift is used over a similarity matrix to identify the possible trajectory followed by the robot and improve the loop closure detection. Our method is tested in a GPS-denied outdoor environment using image sequences collected using a handheld camera and a Parrot ArDrone 2.0.

No description available.

京都大学 / 新制・課程博士 / 博士(情報学) / 甲第25440号 / 情博第878号 / 新制||情||147(附属図書館) / 京都大学大学院情報学研究科システム科学専攻 / (主査)教授 田中 利幸, 教授 下平 英寿, 教授 山下 信雄 / 学位規則第4条第1項該当 / Doctor of Informatics / Kyoto University / DFAM

No description available.

Nous abordons le problème de la mise en correspondance de points dans des images pour calculer la pose d'une caméra par l'algorithme Perspective-n-Point (PnP). Nous calculons la carte 3D, c'est-à-dire les coordonnées 3D et les caractéristiques visuelles de quelques points dans l'environnement grâce à une procédure d'apprentissage hors ligne utilisant un ensemble d'images d'apprentissage. Étant donné une nouvelle image nous utilisons PnP à partir des coordonnées 2D dans l'image de points 3D détectés à l'aide de la carte 3D. Pendant la phase d'apprentissage nous groupons les descripteurs SIFT extraits des images d'apprentissage pour obtenir des collections de positions 2D dans ces images de quelques-uns des points 3D dans l'environnement. Le calcul de SFM (Structure From Motion) est effectué pour obtenir les coordonnées des points correspondants 3D. Pendant la phase de test, les descripteurs SIFT associés aux points 2D projection d'un point 3D de la carte sont utilisés pour reconnaître le point 3D dans une image donnée. Le cadre de travail est semblable à celui des mots visuels utilisés dans différents domaines de la vision par ordinateur. Pendant l'apprentissage, la formation des mots visuelle est effectuée via l'identification de groupes et pendant les tests des points 3D sont identifiés grâce à la reconnaissance des mots visuels. Nous menons des expériences avec des méthodes de formation différentes (k-means et mean-shift) et proposons un nouveau schéma pour la formation des mots visuels pour la phase d'apprentissage. Nous utilisons différentes règles de mise en correspondance, y compris quelques-unes des méthodes standards de classification supervisée pour effectuer la reconnaissance des mots visuels pendant la phase de test. Nous évaluons ces différentes stratégies dans les deux étapes. Afin d'assurer la robustesse aux variations de pose entre images d'apprentissage et images de test, nous explorons différentes façons d'intégrer les descripteurs SIFT extraits de vues synthétiques générées à partir des images d'apprentissage. Nous proposons également une stratégie d'accélération exacte pour l'algorithme mean-shift / We address the problem of establishing point correspondences in images for computing camera pose through Perspective-n-Point (PnP) algorithm. We compute the 3D map i.e. 3D coordinates and visual characteristics of some of the points in the environment through an offline training stage using a set of training images. Given a new test image we apply PnP using the 2D coordinates of 3D points in the image detected by using the 3D map. During the training stage we cluster the SIFT descriptors extracted from training images to obtain 2D-tracks of some of the 3D points in the environment. Each 2D-track consists of a set of 2D image coordinates of a single 3D point in different training images. SfM (Structure from Motion) is performed on these 2D-tracks to obtain the coordinates of the corresponding 3D points. During the test stage, the SIFT descriptors associated the 2D-track of a 3D point is used to recognize the 3D point in a given image. The overall process is similar to visual word framework used in different fields of computer vision. During training, visual word formation is performed through clustering and during testing 3D points are identified through visual word recognition. We experiment with different clustering schemes (k-means and mean-shift) and propose a novel scheme for visual word formation for training stage. We use different matching rules including some of the popular supervised pattern classification methods to perform visual word recognition during test stage. We evaluate these various matching strategies in both stages. In order to achieve robustness against pose variation between train and test images, we explore different ways of incorporating SIFT descriptors extracted from synthetic views generated from the training images. We also propose an exact acceleration strategy for mean-shift computation

Magnetic source magnetic resonance imaging (msMRI) is an fMRI technique that has been under development for direct detection of neuronal magnetic fields to map brain activity and has been shown to be experimentally detectable using conventional means, but there is debate on the detection of the msMRI signal since it can be only a 0.2% change. Detection of its temporal characteristics has yet to be reported and may strengthen the case for msMRI detection. The temporal characteristics of the detected msMRI signal were examined in this work, but it was found that the sensitivity of conventional analysis techniques are low within the context of msMRI, preventing consistent msMRI signal detection and analysis of its temporal characteristics. Examination of blood oxygen level dependent (BOLD) contrast contamination and application of mean-shift clustering (MSC) to fMRI analysis were performed to look into the possibility of improving the low sensitivity. fMRI analysis is commonly performed with cross correlation analysis (CCA) and techniques based on the General Linear Model (GLM), but both CCA and GLM techniques typically perform calculations on a per-voxel basis and do not consider relationships neighboring voxels may have. MSC is a technique to consider for this purpose and shows improved activation detection for both simulated and real BOLD fMRI data. To consider the issue of BOLD contamination, the hemodynamic response over time was examined using repeated median nerve stimulation. On average, the results show the BOLD signal is not detectable after the second fMRI run. The results are consistent with previous hemodynamic habituation effect studies with other types of stimulation, but they do not completely agree with findings of evoked potential studies. Overall, this work shows that the low detection sensitivity may be able to be addressed with the purpose of furthering msMRI research.

In this master thesis, the problem of tracking groups using an image sequence dataset is examined. Target tracking can be defined as the problem of estimating a target's state given prior knowledge about its motion and some sensor measurements related to the target's state. A popular method for target tracking is e.g. the Kalman filter. However, the Kalman filter is insufficient when there are multiple targets in the scene. Consequently, alternative multitarget tracking methods must be applied along with methods for estimating the number of targets in the scene. Multitarget tracking can however be difficult when there are many unresolved targets, e.g. associating observations with targets in dense crowds. A viable simplification is group target tracking, keeping track of groups rather than individual targets. Furthermore, group target tracking is preferred when the user wants to know the motion and extension of a group in e.g. evacuation scenarios. To solve the problem of group target tracking in video surveillance, a combination of GM-PHD filtering and mean shift clustering is proposed. The GM-PHD filter is an approximation of Bayes multitarget filter. Pedestrian detections converted into flat world coordinates from the image dataset are used as input to the filter. The output of the GM-PHD filter consists of Gaussian mixture components with corresponding mean state vectors. The components are divided into groups by using mean shift clustering. An estimate of the number of members and group shape is presented for each group. The method is evaluated using both single camera measurements and two cameras partly surveilling the same area. The results are promising and present a nice visual representation of the groups' characteristics. However, using two cameras gives no improvement in performance, probably due to differences in detections between the two cameras, e.g. a single pedestrian can be observed being at two positions several meters apart making it difficult to determine if it is a single pedestrian or multiple pedestrians.

This thesis presents a robust real-time active tracking system with a pan-tilt camera. The proposed visual servo framework is able to track a deformed object and maintain the target always inside the field of view. For the image processing, an efficient template matching and searching method using the mean-shift theory is developed. The robustness is achieved by appending the ratio histogram, a kernel function, and the template update to the framework when the target is deformed. Then the pan-tilt unit turns towards the target and keeps the target inside the field of view of the camera by feeding back the position information to a Kalman filter. Experimental results show that the presented scheme works successfully when the target is vague or concealed or deformed. The visual tracking task can also be accomplished even when a similar object crosses over the target. In addition, the refreshing rate can be up to 60 frames per second.