The computer vision literature describes many methods to perform obstacle detection and avoidance for autonomous or semi-autonomous vehicles. Methods may be broadly categorized into field-based techniques and feature-...
详细信息
ISBN:
(纸本)0819411922
The computer vision literature describes many methods to perform obstacle detection and avoidance for autonomous or semi-autonomous vehicles. Methods may be broadly categorized into field-based techniques and feature-based techniques. Field-based techniques have the advantage of regular computational structure at every pixel throughout the image plane. Feature-based techniques are much more data driven in that computational complexity increases dramatically in regions of the image populated by features. It is widely believed that to run computer vision algorithms in real time a parallel architecture is necessary. Field-based techniques lend themselves to easy parallelization due to their regular computational needs. However, we have found that field-based methods are sensitive to noise and have traditionally been difficult to generalize to arbitrary vehicle motion. Therefore, we have sought techniques to parallelize feature-based methods. This paper describes the computational needs of a parallel feature-based range-estimation method developed by NASA Ames. Issues of processing-element performance, load balancing, and data-flow bandwidth are addressed along with a performance review of two architectures on which the feature-based method has been implemented.
The Australian Defence Science and Technology Organization is developing a single-platform sensorfusion testbed based around an experimental X-band generic pulse Doppler radar. Initial research will examine real-time...
详细信息
ISBN:
(纸本)0818641207
The Australian Defence Science and Technology Organization is developing a single-platform sensorfusion testbed based around an experimental X-band generic pulse Doppler radar. Initial research will examine real-time fusion of amplitude monopulse radar azimuth and elevation and video position estimates and the tracking ability of the combined sensor system. The addition of further signal processing and sensors will allow experimental verification of a variety of sensorfusion and management algorithms. Examples of preliminary data are shown and the continuing development of the test-bed and its applications are discussed.
The Australian Defence Science and Technology Organisation is developing a single-platform sensorfusion test-bed based around an experimental X-band generic pulse Doppler radar. Initial research will examine real-tim...
详细信息
The Australian Defence Science and Technology Organisation is developing a single-platform sensorfusion test-bed based around an experimental X-band generic pulse Doppler radar. Initial research will examine real-time fusion of amplitude monopulse radar azimuth and elevation and video position estimates and the tracking ability of the combined sensor system. The addition of further signal processing and sensors will allow experimental verification of a variety of sensorfusion and management algorithms. Examples of preliminary data are shown and the continuing development of the test-bed and its applications are discussed.< >
Tree classifiers assign an observation to a class through a series of binary questions. This form of classification is very fast and easy to interpret. However, tree classifiers constructed using standard techniques, ...
详细信息
Tree classifiers assign an observation to a class through a series of binary questions. This form of classification is very fast and easy to interpret. However, tree classifiers constructed using standard techniques, such as CART (classification and regression trees), have difficulties with multi-modal problems like the parity problem. In particular. CART produces a very inefficient tree for this class of problems, which can occur in a number of important applications. This paper examines the problems with CART and then presents a solution that yields trees that use the optimal multi-variate split at each node.< >
Interest in sensor data integration, or data fusion, has grown tremendously in the last five years. The Department of Defense has placed it on its list of critical technologies. The Joint Directors of Laboratories has...
详细信息
The design of image analysis systems aimed at recognition of multiple deformable objects require integration of a variety of modules. These systems (i) are first decomposed into a set of interdependent modules where e...
详细信息
The design of image analysis systems aimed at recognition of multiple deformable objects require integration of a variety of modules. These systems (i) are first decomposed into a set of interdependent modules where each module is associated with a model-based objective, and (ii) are then formed by integrating these modules within a unifying framework. Previous approaches have been limited in providing a systematic and uniform treatment of the design problem and preserving the coexisting and separate nature of the modules' objectives. In Bozma1 a framework which overcomes these limitations has been formulated using game-theoretic concepts. The contribution of this paper is to present an image analysis system which is developed systematically using the game-theoretic framework, and which is aimed at recognizing multiple deformable objects. A secondary contribution is that the models used in designing the individual modules constitute extensions of earlier efforts. Our experiments with several images, including ones from the medical domain, demonstrate the power of this system in applications such as medical image analysis.
The problem of tasking sensors and tracking multiple objects using multiple sensors which are distributed both in space and on the ground is addressed for both commercial and military applications. The data associatio...
详细信息
The problem of tasking sensors and tracking multiple objects using multiple sensors which are distributed both in space and on the ground is addressed for both commercial and military applications. The data association problem is considered in a broader sense in which the assignment of an asset to an object under consideration has been added as an additional function to a traditional tracking and data association system. Suggested learning network architectures and algorithms for potentially enhancing the performance of traditional techniques are discussed and conceptualized.< >
A novel signal representation based on the theory of evidential reasoning was developed. This model, referred to as an evidential signal, represents many competing or overlapping discrete signal hypotheses within a si...
详细信息
A novel signal representation based on the theory of evidential reasoning was developed. This model, referred to as an evidential signal, represents many competing or overlapping discrete signal hypotheses within a single entity. Evidential signals may be processed much like an ordinary discrete signal and are best suited to Bayesian-like hypothesis testing in the absence of prior knowledge. Sum, product, and correlation operations on evidential signals and, in addition, a fusion operation that has no direct analogy in standard signal processing are defined. Simulations demonstrate the application of these principles to a multisensor detection problem.< >
The study of perception, an essential function for autonomous robotics, has led to new developments in the fields of modeling, stochastic data processing, and control. The authors describe a surveillance robot dedicat...
详细信息
ISBN:
(纸本)0879426888
The study of perception, an essential function for autonomous robotics, has led to new developments in the fields of modeling, stochastic data processing, and control. The authors describe a surveillance robot dedicated to heterogeneous data fusion algorithms' implementation and real-size real-time testing, which has become necessary to validate these emerging techniques in a realistic environment. This robot holds various sensors mounted on rotational units and provides a multi-degree-of-freedom command. The design includes original distributed hardware and software architectures encompassing the specific data processing, fusion, and feedback control. Modularity and a user-friendly interface allow easy design of perception control applications.
The volume contains 107 conference papers. The main topics covered include image processing, electrooptical sensor signal processing, nonlinear systems, knowledge-based signal processing, speech applications, signal p...
详细信息
The volume contains 107 conference papers. The main topics covered include image processing, electrooptical sensor signal processing, nonlinear systems, knowledge-based signal processing, speech applications, signal processing and coding for digital storage systems, multiprocessor architectures, adaptive filters, array signal processing, computer vision, residue number systems, fault tolerance and testing, signal processing for digital facsimile, VLSI issues in algorithms and computations, signal subspace methods, neural nets in signal processing, evolutionary programming, DSP implementations, and applications of time-frequency signal representations.
暂无评论