Centre dedocumentation
scientifique

Adaptive correlation filters with long-term and short-term memory for object tracking / Chao Ma in International journal of computer vision, vol 126 n° 8 (August 2018)  

Public [article]  inInternational journal of computer vision > vol 126 n° 8 (August 2018) . - pp 771 - 796 Titre : Adaptive correlation filters with long-term and short-term memory for object tracking Type de document : Article/Communication Auteurs : Chao Ma, Auteur ; Jia-Bin Huang, Auteur ; Xiaokang Yang, Auteur ; Ming-Hsuan Yang, Auteur Année de publication : 2018 Article en page(s) : pp 771 - 796 Note générale : Bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Traitement d'image [Termes descripteurs IGN] détection d'objet [Termes descripteurs IGN] filtre adaptatif [Termes descripteurs IGN] méthode fondée sur le noyau [Termes descripteurs IGN] méthode robuste [Termes descripteurs IGN] poursuite de cible Résumé : (Auteur) Object tracking is challenging as target objects often undergo drastic appearance changes over time. Recently, adaptive correlation filters have been successfully applied to object tracking. However, tracking algorithms relying on highly adaptive correlation filters are prone to drift due to noisy updates. Moreover, as these algorithms do not maintain long-term memory of target appearance, they cannot recover from tracking failures caused by heavy occlusion or target disappearance in the camera view. In this paper, we propose to learn multiple adaptive correlation filters with both long-term and short-term memory of target appearance for robust object tracking. First, we learn a kernelized correlation filter with an aggressive learning rate for locating target objects precisely. We take into account the appropriate size of surrounding context and the feature representations. Second, we learn a correlation filter over a feature pyramid centered at the estimated target position for predicting scale changes. Third, we learn a complementary correlation filter with a conservative learning rate to maintain long-term memory of target appearance. We use the output responses of this long-term filter to determine if tracking failure occurs. In the case of tracking failures, we apply an incrementally learned detector to recover the target position in a sliding window fashion. Extensive experimental results on large-scale benchmark datasets demonstrate that the proposed algorithm performs favorably against the state-of-the-art methods in terms of efficiency, accuracy, and robustness. Numéro de notice : A2018-414 Thématique : IMAGERIE Nature : Article nature-HAL : ArtAvecCL-RevueIntern DOI : 10.1007/s11263-018-1076-4 date de publication en ligne : 16/03/2018 En ligne : https://doi.org/10.1007/s11263-018-1076-4 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=90897 [article]

Object tracking with robotic total stations : current technologies and improvements based on image data / Matthias Ehrhart in Journal of applied geodesy, vol 11 n° 3 (September 2017)  

Public [article]  inJournal of applied geodesy > vol 11 n° 3 (September 2017) . - pp 131 - 142 Titre : Object tracking with robotic total stations : current technologies and improvements based on image data Type de document : Article/Communication Auteurs : Matthias Ehrhart, Auteur ; Werner Lienhart, Auteur Année de publication : 2017 Article en page(s) : pp 131 - 142 Note générale : Bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Topographie [Termes descripteurs IGN] instrumentation Leica [Termes descripteurs IGN] instrumentation Topcon [Termes descripteurs IGN] instrumentation Trimble [Termes descripteurs IGN] positionnement par GNSS [Termes descripteurs IGN] poursuite de cible [Termes descripteurs IGN] prisme (optique) [Termes descripteurs IGN] tachéomètre électronique robotisé [Termes descripteurs IGN] vidéotachéomètre Résumé : (Auteur) The importance of automated prism tracking is increasingly triggered by the rising automation of total station measurements in machine control, monitoring and one-person operation. In this article we summarize and explain the different techniques that are used to coarsely search a prism, to precisely aim at a prism, and to identify whether the correct prism is tracked. Along with the state-of-the-art review, we discuss and experimentally evaluate possible improvements based on the image data of an additional wide-angle camera which is available for many total stations today. In cases in which the total station’s fine aiming module loses the prism, the tracked object may still be visible to the wide-angle camera because of its larger field of view. The theodolite angles towards the target can then be derived from its image coordinates which facilitates a fast reacquisition of the prism. In experimental measurements we demonstrate that our image-based approach for the coarse target search is 4 to 10-times faster than conventional approaches. Numéro de notice : A2017-568 Thématique : POSITIONNEMENT Nature : Article DOI : 10.1515/jag-2016-0043 En ligne : https://doi.org/10.1515/jag-2016-0043 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=86688 [article]

Motion priors based on goals hierarchies in pedestrian tracking applications / Francisco Madrigal in Machine Vision and Applications, vol 28 n° 3-4 (May 2017)  

Public [article]  inMachine Vision and Applications > vol 28 n° 3-4 (May 2017) . - pp 341 - 359 Titre : Motion priors based on goals hierarchies in pedestrian tracking applications Type de document : Article/Communication Auteurs : Francisco Madrigal, Auteur ; Jean-Bernard Hayet, Auteur Année de publication : 2017 Article en page(s) : pp 341 - 359 Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Traitement d'image optique [Termes descripteurs IGN] apprentissage automatique [Termes descripteurs IGN] carrefour [Termes descripteurs IGN] compréhension de l'image [Termes descripteurs IGN] image vidéo [Termes descripteurs IGN] modèle de simulation [Termes descripteurs IGN] position [Termes descripteurs IGN] poursuite de cible [Termes descripteurs IGN] prévision [Termes descripteurs IGN] réalité de terrain [Termes descripteurs IGN] séquence d'images Résumé : (auteur) In this paper, the problem of automated scene understanding by tracking and predicting paths for multiple humans is tackled, with a new methodology using data from a single, fixed camera monitoring the environment. Our main idea is to build goal-oriented prior motion models that could drive both the tracking and path prediction algorithms, based on a coarse-to-fine modeling of the target goal. To implement this idea, we use a dataset of training video sequences with associated ground-truth trajectories and from which we extract hierarchically a set of key locations. These key locations may correspond to exit/entrance zones in the observed scene, or to crossroads where trajectories have often abrupt changes of direction. A simple heuristic allows us to make piecewise associations of the ground-truth trajectories to the key locations, and we use these data to learn one statistical motion model per key location, based on the variations of the trajectories in the training data and on a regularizing prior over the models spatial variations. We illustrate how to use these motion priors within an interacting multiple model scheme for target tracking and path prediction, and we finally evaluate this methodology with experiments on common datasets for tracking algorithms comparison. Numéro de notice : A2017-325 Thématique : IMAGERIE Nature : Article En ligne : http://doi.org/10.1007/s00138-017-0832-8 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=85384 [article]

Probabilistic multi-person localisation and tracking / Tobias Klinger (2016) 

Public Titre : Probabilistic multi-person localisation and tracking Type de document : Thèse/HDR Auteurs : Tobias Klinger, Auteur ; Ingo Neumann, Directeur de thèse Editeur : Munich : Bayerische Akademie der Wissenschaften Année de publication : 2016 Collection : DGK - C, ISSN 0065-5325 num. 787 Importance : 125 p. Format : 21 x 30 cm ISBN/ISSN/EAN : 978-3-7696-5199-7 Note générale : bibliographie PhD Dissertation Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Traitement d'image optique [Termes descripteurs IGN] analyse de sensibilité [Termes descripteurs IGN] analyse multicritère [Termes descripteurs IGN] cible mobile [Termes descripteurs IGN] classification [Termes descripteurs IGN] détection de piéton [Termes descripteurs IGN] géolocalisation [Termes descripteurs IGN] image isolée [Termes descripteurs IGN] modèle stochastique [Termes descripteurs IGN] piéton [Termes descripteurs IGN] poursuite de cible [Termes descripteurs IGN] programmation linéaire [Termes descripteurs IGN] séquence d'images [Termes descripteurs IGN] similitude [Termes descripteurs IGN] surveillance Résumé : (auteur) This dissertation investigates the problem of localising multiple persons in image sequences, while, at the same time, establishing temporal correspondences between single-frame locations. The aim of this work is the improvement of the reliability and precision of the generated trajectories, which is addressed by the formulation and investigation of a joint probabilistic model for the recursive filtering of the estimated positions. The trajectories are estimated in a common 3D object coordinate system, which was previously almost exclusively done in 2D. Note de contenu : 1. Introduction 1.1. Motivation 1.2. Research objectives and contributions 1.3. Outline of the dissertation 2. Basics 2.1. Probabilistic modelling 2.2. Recursive Bayesian estimation 2.3. Gaussian Process Regression 3. Related work 3.1. Tracking approaches 3.2. Observations 3.3. Temporal modelling 3.4. Data association 3.5. Discussion 4. A new probabilistic approach for multi-person localisation and tracking 4.1. Problem statement via Dynamic Bayesian Network 4.2. Observations 4.3. Temporal model 4.4. data association 4.5. Recursive estimation 4.6. Discussion 5. Experiments 5.1. Datasets and evaluation criteria 5.2. Sensitivity study and training 5.3. Model validation by ablation of its components 5.4. Multi-person localisation and tracking evaluation 6. Discussion of the results 6.1. Method evaluation 6.2. Evaluation of the trajectories 7. Conclusions and future work Numéro de notice : 19793 Thématique : IMAGERIE/INFORMATIQUE Nature : Thèse étrangère Note de thèse : PhD Dissertation : : Stuttgart : 2016 Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=85037

Documents numériques

peut être téléchargé Probabilistic multi-person localisation and tracking Adobe Acrobat PDF

Tracking 3D moving objects based on GPS/IMU navigation solution, laser scanner point cloud and GIS data / Siavash Hosseinyalamdary in ISPRS International journal of geo-information, vol 4 n°3 (September 2015)  

Public [article]  inISPRS International journal of geo-information > vol 4 n°3 (September 2015) . - pp 1301 - 1316 Titre : Tracking 3D moving objects based on GPS/IMU navigation solution, laser scanner point cloud and GIS data Type de document : Article/Communication Auteurs : Siavash Hosseinyalamdary, Auteur ; Yashar Balazadegan, Auteur ; Charles K. Toth, Auteur Année de publication : 2015 Article en page(s) : pp 1301 - 1316 Note générale : bibliographie Langues : Anglais (eng) Descripteur : [Vedettes matières IGN] Applications photogrammétriques [Termes descripteurs IGN] détection d'objet [Termes descripteurs IGN] données localisées 3D [Termes descripteurs IGN] filtre de Kalman [Termes descripteurs IGN] objet géographique 3D [Termes descripteurs IGN] objet mobile [Termes descripteurs IGN] poursuite de cible [Termes descripteurs IGN] surveillance routière [Termes descripteurs IGN] trafic routier Résumé : (auteur) Monitoring vehicular road traffic is a key component of any autonomous driving platform. Detecting moving objects, and tracking them, is crucial to navigating around objects and predicting their locations and trajectories. Laser sensors provide an excellent observation of the area around vehicles, but the point cloud of objects may be noisy, occluded, and prone to different errors. Consequently, object tracking is an open problem, especially for low-quality point clouds. This paper describes a pipeline to integrate various sensor data and prior information, such as a Geospatial Information System (GIS) map, to segment and track moving objects in a scene. We show that even a low-quality GIS map, such as OpenStreetMap (OSM), can improve the tracking accuracy, as well as decrease processing time. A bank of Kalman filters is used to track moving objects in a scene. In addition, we apply non-holonomic constraint to provide a better orientation estimation of moving objects. The results show that moving objects can be correctly detected, and accurately tracked, over time, based on modest quality Light Detection And Ranging (LiDAR) data, a coarse GIS map, and a fairly accurate Global Positioning System (GPS) and Inertial Measurement Unit (IMU) navigation solution. Numéro de notice : A2015-711 Thématique : IMAGERIE Nature : Article nature-HAL : ArtAvecCL-RevueIntern En ligne : http://www.mdpi.com/2220-9964/4/3/1301 Format de la ressource électronique : URL article Permalink : https://documentation.ensg.eu/index.php?lvl=notice_display&id=78348 [article]

Tracking-Learning-Detection / Zdenek Kalal in IEEE Transactions on Pattern Analysis and Machine Intelligence, PAMI, vol 34 n° 7 (July 2012)

Permalink

Indoor pedestrian navigation using foot-mounted IMU and portable ultrasound range sensors / Gabriel Girard in Sensors, vol 11 n° 8 (August 2011) 

Permalink

Photogrammetric image analysis PIA 11, Munich, Germany, October 5-7, 2011 / Uwe Stilla (2011)  

Permalink