The availability of video imagery through reduction in sensor size, cost, and power has enabled an explosion of collection opportunities. With the increased amount of imagery there is a need to understand the usefulne...
详细信息
The availability of video imagery through reduction in sensor size, cost, and power has enabled an explosion of collection opportunities. With the increased amount of imagery there is a need to understand the usefulness of video for applications such as Activity-Based Intelligence (ABI), situation understanding, and event-based processing. In this paper, we explore some of the emerging developments in video observations with a focus on cloud technology. Cloud technology supports integration of multiple algorithms, storage of large data sets, indexing over multimedia, and workflow opportunities between humans and machines. We highlight multiple tools such as GeoFlix™, Application Knowledge Interface To algorithms (AKITA™), and Intelligence Preparation of the Operational Environment (IPOE) using the Ozone Widget Framework (OWF) for permissive surveillance, data to decisions, and information fusion. These tools enable data analytics, algorithm comparison, and user-defined visualizations. An example is presented for target localization and tracking through planned video observations.
Most of the mobile applications require efficient and precise computation of the device pose, and almost every mobile device has inertial sensors already equipped together with a camera. This fact makes sensor fusion ...
详细信息
ISBN:
(纸本)9781479957521
Most of the mobile applications require efficient and precise computation of the device pose, and almost every mobile device has inertial sensors already equipped together with a camera. This fact makes sensor fusion quite attractive for increasing efficiency during pose tracking. However, the state-of-the-art fusionalgorithms have a major shortcoming: lack of well-defined uncertainty introduced to the system during the prediction stage of the fusion filters. Such a drawback results in determining covariances heuristically, and hence, requirement for data-dependent tuning to achieve high performance or even convergence of these filters. In this paper, we propose an inertially-aided visual odometry system that requires neither heuristics nor parameter tuning;computation of the required uncertainties on all the estimated variables are obtained after minimum number of assumptions. Moreover, the proposed system simultaneously estimates the metric scale of the pose computed from a monocular image stream. The experimental results indicate that the proposed scale estimation outperforms the state-of-the-art methods, whereas the pose estimation step yields quite acceptable results in real-time on resource constrained systems.
We present a real-time, full-frame, multi-target Wide Area Motion Imagery (WAMI) tracking system that utilizes distributed processing to handle high data rates while maintaining high track quality. The proposed archit...
详细信息
ISBN:
(纸本)9781479949847
We present a real-time, full-frame, multi-target Wide Area Motion Imagery (WAMI) tracking system that utilizes distributed processing to handle high data rates while maintaining high track quality. The proposed architecture processes the WAMI data as a series of geospatial tiles and implements both process- and thread-level parallelism across multiple compute nodes. Each tile is processed independently, from decoding the image through generating tracks that are finally merged across all tiles by an inter-tile linker (ITL) module. A high performance PostgreSQL database with GIS extensions is used to control the flow of intermediate data between each tracking process. High quality tracks are produced efficiently due to robust, effective algorithmic modules including: multi-frame moving object detection and track initialization; tracking based on the fusion of motion and appearance with a goal of very pure tracks; and online track linking based on multiple features. In addition, we have configured a high-performance compute cluster using high density blade servers, Infiniband networking, and an HPC filesystem. The compute cluster enables full-frame, state-of-the-art tracking of vehicles or dismounts at the WAMI sensor's native 1.25Hz frame-rate, while only taking 7u of rack space and providing 210 megapixels/second throughput.
The proceedings contain 44 papers. The special focus in this conference is on Neural Networks. The topics include: Identifying emergent dynamical structures in network models;experimental guidelines for semantic-based...
ISBN:
(纸本)9783319041285
The proceedings contain 44 papers. The special focus in this conference is on Neural Networks. The topics include: Identifying emergent dynamical structures in network models;experimental guidelines for semantic-based regularization;a preliminary study on transductive extreme learning machines;avoiding the cluster hypothesis in SV classification of partially labeled data;learning capabilities of ELM-trained time-varying neural networks;a quality-driven ensemble approach to automatic model selection in clustering;an adaptive reference point approach to efficiently search large chemical databases;genetic art in perspective;proportionate algorithms for blind source separation;pupillometric study of the dysregulation of the autonomous nervous system by SVM networks;a memristor circuit using basic elements with memory capability;effects of pruning on phase-coding and storage capacity of a spiking network;predictive analysis of the seismicity level at campi flegrei volcano using a data-driven approach;robot localization by echo state networks using RSS;smart home task and energy resource scheduling based on nonlinear programming;datafusion using a factor graph for ship tracking in harbour scenarios;reinforcement learning for automated financial trading;a collaborative filtering recommender exploiting a SOM network;SVM tree for personalized transductive learning in bioinformatics classification problems;multi-country mortality analysis using self organizing maps;a fuzzy decision support system for the environmental risk assessment of genetically modified organisms;recent approaches in handwriting recognition with Markovian modelling and recurrent neural networks;the effects of hand gestures on psychosocial perception;the influence of positive and negative emotions on physiological responses and memory task scores;mood effects on the decoding of emotional voices;the ascending reticular activating system;conversational entrainment in the use of discourse markers;language an
Localisation is one of the most important applications for wireless sensor networks since the locations of the sensor nodes are critical to both network operations and most application level tasks. Numerous techniques...
详细信息
ISBN:
(纸本)9781479935130
Localisation is one of the most important applications for wireless sensor networks since the locations of the sensor nodes are critical to both network operations and most application level tasks. Numerous techniques for localisation of sensor nodes that make use of the Received Signal Strength Indicator (RSSI) have been proposed because of the simplicity and low cost of implementation. However, most of the research thus far has regarded the RSSI technology as unsuitable for accurate localisation due to the limited accuracy inherent to the current ranging models. These models make the assumption that the antenna radiation pattern is omnidirectional in order to simplify the complexity of the algorithms. In this study, an accurate and efficient localisation method that makes use of an improved RSSI distance estimation model by including the antenna radiation pattern as well as nodes orientations is presented. Mathematical models for distance estimation, cost function and gradient of cost function, that can be used in a distributed localisation algorithm, are developed. This study also introduces a sensor datafusion approach, combining accelerometer data, RSSI, antenna radiation pattern and node orientation to reduce the computation complexity during the tracking phase. The proposed algorithm is implemented in Matlab. Simulation results show that the proposed approach increases the accuracy of existing methods using RSSI by up to 59%.
This work addresses the problem of extracting semantics associated with multiple, cooperatively managed motion imagery sensors to support indexing and search of large imagery collections. The extracted semantics relat...
详细信息
This work addresses the problem of extracting semantics associated with multiple, cooperatively managed motion imagery sensors to support indexing and search of large imagery collections. The extracted semantics relate to the motion and identity of vehicles within a scene, viewed from aircraft and the ground. Semantic extraction required three steps: Video Moving target Indication (VMTI), imagery fusion, and object recognition. VMTI used a previously published algorithm, with some novel modifications allowing detection and tracking in low frame rate, Wide Area Motion Imagery (WAMI), and Full Motion Video (FMV). Following this, the data from multiple sensors were fused to identify a highest resolution image, corresponding to each moving object. A final recognition stage attempted to fit each delineated object to a database of 3D models to determine its type. A proof-of-concept has been developed to allow processing of imagery collected during a recent experiment using a state of the art airborne surveillance sensor providing WAMI, with coincident narrower-area FMV sensors and simultaneous collection by a ground-based camera. An indication of the potential utility of the system was obtained using ground-truthed examples.
Without any dedicated radiation, Passive radars are able to detect and localize air targets and especially low altitude targets. The actual passive coverage for these low altitude targets is generally sufficient for g...
详细信息
Without any dedicated radiation, Passive radars are able to detect and localize air targets and especially low altitude targets. The actual passive coverage for these low altitude targets is generally sufficient for gap-fillers applications such as the continuous survey of sensitive areas or huge cities under the following condition: the passive system have to evaluate the target altitude to ensure that the target is a threat or at least an intruder. In order to fulfill the low altitude coverage as well as the required altitude accuracy, the main passive component has to be based on DVB-T (Digital Video Broadcaster-Terrestrial) . After a short comparison of the two main approaches (triangulation versus direct site measurement) for this altitude estimation, an example of a 3D-DVB-T receiving component will be presented. This paper will then focus on the experimental results achieved using such a small 3D-DVB-T component. The potential improvement of a fusion with other sub-systems such as 2DDVBT component and/or FM component will be also illustrated using experimental data. These works and experimentation were funded by the French MoD (DGA) and were conducted in collaboration with THALES.
暂无评论