Proceedings of ICSIPNN '94. International Conference on Speech, Image Processing and Neural Networks
2. The Basis Of The Method In this paper we present a texture analysis and classification method ... more 2. The Basis Of The Method In this paper we present a texture analysis and classification method based on the computation of highorder statistics of gray-levels or selected features. The computation of these high-order statistics can be performed in parallel in non-overlapping windows that cover the full image. Classification is performed by neural networks. In this extended summary results are presented for back-propagation networks only, but several other types of neural nets were also tested.
Proceedings 2000 ICRA. Millennium Conference. IEEE International Conference on Robotics and Automation. Symposia Proceedings (Cat. No.00CH37065)
In this paper we describe a focusing-by-vergence system combining a ow based vergence process wit... more In this paper we describe a focusing-by-vergence system combining a ow based vergence process with an o-line pre-calibration of the focusing odometry. The vergence process is controlled by the retinal optical ow disparity and the target depth velocity is directly obtained from this disparity. The relationship between the target depth (focused target distance) and the focusing odometry of the motorized lens is obtained by an o-line focusing calibration, using a bivariate polynomial to model de relationship between these parameters. The focusing system is velocity controlled and the focusing velocity is dependent on the target focusing depth velocity and the focal distance velocity. The in uence of the focal distance velocity that arise due to the focusing process itself is compensated by the adjustment of the zoom parameter based on a constant magni cation approach, or through the modeling of the focal distance as a function of the zoom and focus lens settings.
Object recognition supported by user interaction for service robots
The estimation of camera egomotion is a well established problem in computer vision. Many approac... more The estimation of camera egomotion is a well established problem in computer vision. Many approaches have been proposed based on both the discrete and the differential epipolar constraint. The discrete case is mainly used in self-calibrated stereoscopic systems, whereas the differential case deals with an unique moving camera. This article surveys several methods for mobile robot egomotion estimation covering more than 0.5 million samples using synthetic data. Results from real data are also given. The surveyed algorithms have been programmed and are available on the Internet: http://eia.udg.es/~armangue/research.
Proceedings IECON '91: 1991 International Conference on Industrial Electronics, Control and Instrumentation
The authors address the geometric modeling aspects related to the use of a pair of video cameras ... more The authors address the geometric modeling aspects related to the use of a pair of video cameras mounted on a six-degree-of-freedom manipulator in a hand-eye configuration. The information obtained by these sensors will be used to generate the depth map of the environment where the manipulator works. The procedure for determining the image formation parameters is known by camera calibration,
One interesting feature of the human visual system is the topological transformation of the retin... more One interesting feature of the human visual system is the topological transformation of the retinal image into its cortical projection. The excitation of the cortex can be approximated by a log-polar mapping of the eye's retinal image. In this tutorial paper we describe the log-polar mapping and its main properties.
Proceedings 1999 IEEE/RSJ International Conference on Intelligent Robots and Systems. Human and Environment Friendly Robots with High Intelligence and Emotional Quotients (Cat. No.99CH36289)
In this paper a system for tracking multiple targets in 3D is described. The system is made up of... more In this paper a system for tracking multiple targets in 3D is described. The system is made up of two pan-and-tilt units that are attached to the extremities of a rotating arm. This configuration has several advantages and can deal with several specific instances of tracking more than one target. A control strategy that guarantees equal target disparities in both images whenever targets are seen by both cameras is presented. This has advantages for segmentation and trajectory reconstruction. Target images are simultaneously visible in the cameras, enabling the recovery of the targets 3D trajectory. It is also shown that mutual occlusion occurs in a well-defined configuration and can therefore be dealt with.
Proceedings. Fourteenth International Conference on Pattern Recognition (Cat. No.98EX170)
Surveillance applications have specific features that de-pend upon the type of environment they a... more Surveillance applications have specific features that de-pend upon the type of environment they are designed to op-erate on. In most applications vision is the main sensing modality, even thoughin very specialized applicationsother sensors are used. The scenarios of surveillance ...
Proceedings of 13th International Conference on Pattern Recognition, 1996
A n active vision system has to enable the implementation of reactzve visual processes and of ele... more A n active vision system has to enable the implementation of reactzve visual processes and of elementary visual behaviors .in real tzme. Therefore the control architecture is extremely important. In this paper we discuss a number of issues related with the implementation of a real-ttme control architecture and describe the architecture we are using with camera heads. Another important issue of the operation of active vision binocular heads is their integration into more complex robotic systems. 'The design of the control architecture has to be suited to the integration of the system in other robotic systems. Higher levels of autonomy and integration can be obtained by designing the system architecture based on the concept of purposive behavior. A t the lower levels we consider vision as a sensor a n d integrate it in control systems (both feed-forward and servo loops) and several visual processes are implemented in parallel, computing relevant measures for the control process. A t higher levels the architecture is modeled as a state transition system. Finally we show how this architecture can be used to implement a pursuit behavior using optical pow. Simultaneously vergence control can also be performec! using the same visual processes.
This paper describes the use of an experimental system for active vision to compute time until im... more This paper describes the use of an experimental system for active vision to compute time until impact. The time-until-impact is evaluated by using the principle of the human visual fixation process. The mechanical system enables the positioning of the cameras' support platform by means of three rotational degrees of freedom and also the control of the vergence angle of each camera. The motors used are stepper motors which are controlled by a system specifically designed and built for this purpose. Each camera's lens has three degrees of freedom , corresponding to zoom, focus and aperture. All the control system electronics as well as the system status data acquisition electronics are integrated in a single control unit. This control unit is connected to a system responsible for the active vision system motion generation.
Lecture Notes in Control and Information Sciences, 2000
In this paper we describe how the combination of several vision sensors can be used to track mult... more In this paper we describe how the combination of several vision sensors can be used to track multiple human targets in a typical surveillance situation. The experimental system devised is made up of a xed camera and a binocular active vision system. The xed camera is used to provide the system with peripheral motion detection capabilities. The human activity is monitored by extracting several parameters that are useful for their classi cation. The system enables the creation of a record based on the type of activity. These logs can be selectively accessed and provide images of the humans in speci c areas.
Lecture Notes in Control and Information Sciences, 1997
This article describes an algorithm for detection of obstacles, through the use of low cost sonar... more This article describes an algorithm for detection of obstacles, through the use of low cost sonar sensors, for the navigation of mobile robots. The use of a grid representation for the cartesian space, associated to probabilistic operations on it, allowed the implementation of a fast, simple and robust algorithm for obstacle detection. This algorithm runs always that the mobile robot navigates, and allows to overcome obstacles that appear unexpectedly in the robot's path. The information used to represent the obstacles is based in the information provided by sonar sensors. A stochastic model for the sonar sensors and their activity, associated to the platform motion, are explored by the algorithm. Some experimental results obtained through the use this algorithm are described at the end of the article.
Tracking in 3D with an active vision system depends on the performance of both motor control and ... more Tracking in 3D with an active vision system depends on the performance of both motor control and vision algorithms. Tracking is performed based on different visual behaviors, namely smooth pursuit and vergence control. A major issue in a system performing tracking is its robustness to partial occlusion of the target as well as its robustness to sudden changes of target trajectory. Another important issue is the reconstruction of the 3D trajectory of the target. These issues can only be dealt with if the performance of the algorithms is evaluated. The evaluation of such performances enable the identification of the limits and weaknesses in the system behavior. In this paper we describe the results of the analysis of a binocular tracking system. To perform the evaluation a control framework was used both for the vision algorithms and for the servo-mechanical system. Due to the geometry changes in an active vision system, the problem of defining and generating system reference inputs has specific features. In this paper we analyze this problem, proposing and justifying a methodology for the definition and generation of such reference inputs. As a result several algorithms were improved and the global performance of the system was also enhanced. This paper proposes a methodology for such an analysis (and resulting enhancements) based on techniques from control theory.
Lecture Notes in Control and Information Sciences, 1998
In this paper we discuss the use of active vision for the development of autonomous systems. Acti... more In this paper we discuss the use of active vision for the development of autonomous systems. Active vision systems are essentially based on biological motivations. Two systems with potential application to surveillance are described. Both systems behave as watchrobots. One of them involves the integration of an active vision system in a mobile platform. The second system can track non-rigid
The robustness of visual behaviors implemented in an active vision sys- tem depends both on the v... more The robustness of visual behaviors implemented in an active vision sys- tem depends both on the vision algorithms and control structure. To improve robustness an evaluation of the performance of the several system algorithms must be done. Performance analysis can be done within the framework of con- trol theory since notions such as stability and controllability can contribute to a
Microprocessors in Robotic and Manufacturing Systems, 1991
Sensor-based robots can increase significantly the scope of applications of robots in manufacturi... more Sensor-based robots can increase significantly the scope of applications of robots in manufacturing namely for flexible assembly. This paper describes the development of a robotic system used for the evaluation of sensory integration techniques and real-time path control strategies. The robotic system has a distributed hierarchical architecture, featuring vision, force, tactile, range and proximity sensors. The approaches for integration of
Object recognition supported by user interaction for service robots
In this paper we describe a real-time gesture recognition algorithm based on contour signature im... more In this paper we describe a real-time gesture recognition algorithm based on contour signature images, generated by collecting a set of contour signatures over time. Each contour signature is generated in each frame by computing the polar coordinates of each point of the contour in relation to the center of mass of that contour. The gesture recognition is reduced to the problem of finding the correspondent contour signature image from a set of previously learned image contour signatures. This is achieved using a Karhunen-Loeve (KL) transform that reduces the set of trained images to a low-dimensional eigenspace, making recognition easier. This method is particularly suitable for the recognition of gestures that can be perfectly described by silhouettes. Since no shape or 3D model is required the proposed method can easily be adapted to recognize several other types of gestures not necessarily performed by humans. The proposed algorithm may be able to extract target gesture motions and classify them at the same time. Preliminary experimental results demonstrate robustness and real-time performance of the proposed algorithm.
Procedings of the British Machine Vision Conference 1993, 1993
A new method for robotics camera calibration based on a linear computation and using a coplanar g... more A new method for robotics camera calibration based on a linear computation and using a coplanar group of calibration points is presented in this paper. This new method suppresses the constraint presented by the RAC calibration model of having an incidence angle between the optical axis and the calibration plane of at least 30 degrees. The parameters are obtained based on a multi-step calculation, and the accuracy of the parameters is strongly increased when used on a iterative calibration. Experimental and simulated analyses results are presented. A comparative analysis between this new method and the RAC method is presented.
In this paper we describe the integration of vision information obtained by a fixed camera and th... more In this paper we describe the integration of vision information obtained by a fixed camera and the information obtained by a binocular active vision system. Both systems acquire images from the same environment. The information redundancy is exploited to achieve a higher degree of robustness in detecting and tracking intruders. Global information about intruders is obtained by the fixed vision system whereas information about a specific intruder (that invades a critical area, for example) is obtained by the binocular system.
Proceedings of ICSIPNN '94. International Conference on Speech, Image Processing and Neural Networks
2. The Basis Of The Method In this paper we present a texture analysis and classification method ... more 2. The Basis Of The Method In this paper we present a texture analysis and classification method based on the computation of highorder statistics of gray-levels or selected features. The computation of these high-order statistics can be performed in parallel in non-overlapping windows that cover the full image. Classification is performed by neural networks. In this extended summary results are presented for back-propagation networks only, but several other types of neural nets were also tested.
Proceedings 2000 ICRA. Millennium Conference. IEEE International Conference on Robotics and Automation. Symposia Proceedings (Cat. No.00CH37065)
In this paper we describe a focusing-by-vergence system combining a ow based vergence process wit... more In this paper we describe a focusing-by-vergence system combining a ow based vergence process with an o-line pre-calibration of the focusing odometry. The vergence process is controlled by the retinal optical ow disparity and the target depth velocity is directly obtained from this disparity. The relationship between the target depth (focused target distance) and the focusing odometry of the motorized lens is obtained by an o-line focusing calibration, using a bivariate polynomial to model de relationship between these parameters. The focusing system is velocity controlled and the focusing velocity is dependent on the target focusing depth velocity and the focal distance velocity. The in uence of the focal distance velocity that arise due to the focusing process itself is compensated by the adjustment of the zoom parameter based on a constant magni cation approach, or through the modeling of the focal distance as a function of the zoom and focus lens settings.
Object recognition supported by user interaction for service robots
The estimation of camera egomotion is a well established problem in computer vision. Many approac... more The estimation of camera egomotion is a well established problem in computer vision. Many approaches have been proposed based on both the discrete and the differential epipolar constraint. The discrete case is mainly used in self-calibrated stereoscopic systems, whereas the differential case deals with an unique moving camera. This article surveys several methods for mobile robot egomotion estimation covering more than 0.5 million samples using synthetic data. Results from real data are also given. The surveyed algorithms have been programmed and are available on the Internet: http://eia.udg.es/~armangue/research.
Proceedings IECON '91: 1991 International Conference on Industrial Electronics, Control and Instrumentation
The authors address the geometric modeling aspects related to the use of a pair of video cameras ... more The authors address the geometric modeling aspects related to the use of a pair of video cameras mounted on a six-degree-of-freedom manipulator in a hand-eye configuration. The information obtained by these sensors will be used to generate the depth map of the environment where the manipulator works. The procedure for determining the image formation parameters is known by camera calibration,
One interesting feature of the human visual system is the topological transformation of the retin... more One interesting feature of the human visual system is the topological transformation of the retinal image into its cortical projection. The excitation of the cortex can be approximated by a log-polar mapping of the eye's retinal image. In this tutorial paper we describe the log-polar mapping and its main properties.
Proceedings 1999 IEEE/RSJ International Conference on Intelligent Robots and Systems. Human and Environment Friendly Robots with High Intelligence and Emotional Quotients (Cat. No.99CH36289)
In this paper a system for tracking multiple targets in 3D is described. The system is made up of... more In this paper a system for tracking multiple targets in 3D is described. The system is made up of two pan-and-tilt units that are attached to the extremities of a rotating arm. This configuration has several advantages and can deal with several specific instances of tracking more than one target. A control strategy that guarantees equal target disparities in both images whenever targets are seen by both cameras is presented. This has advantages for segmentation and trajectory reconstruction. Target images are simultaneously visible in the cameras, enabling the recovery of the targets 3D trajectory. It is also shown that mutual occlusion occurs in a well-defined configuration and can therefore be dealt with.
Proceedings. Fourteenth International Conference on Pattern Recognition (Cat. No.98EX170)
Surveillance applications have specific features that de-pend upon the type of environment they a... more Surveillance applications have specific features that de-pend upon the type of environment they are designed to op-erate on. In most applications vision is the main sensing modality, even thoughin very specialized applicationsother sensors are used. The scenarios of surveillance ...
Proceedings of 13th International Conference on Pattern Recognition, 1996
A n active vision system has to enable the implementation of reactzve visual processes and of ele... more A n active vision system has to enable the implementation of reactzve visual processes and of elementary visual behaviors .in real tzme. Therefore the control architecture is extremely important. In this paper we discuss a number of issues related with the implementation of a real-ttme control architecture and describe the architecture we are using with camera heads. Another important issue of the operation of active vision binocular heads is their integration into more complex robotic systems. 'The design of the control architecture has to be suited to the integration of the system in other robotic systems. Higher levels of autonomy and integration can be obtained by designing the system architecture based on the concept of purposive behavior. A t the lower levels we consider vision as a sensor a n d integrate it in control systems (both feed-forward and servo loops) and several visual processes are implemented in parallel, computing relevant measures for the control process. A t higher levels the architecture is modeled as a state transition system. Finally we show how this architecture can be used to implement a pursuit behavior using optical pow. Simultaneously vergence control can also be performec! using the same visual processes.
This paper describes the use of an experimental system for active vision to compute time until im... more This paper describes the use of an experimental system for active vision to compute time until impact. The time-until-impact is evaluated by using the principle of the human visual fixation process. The mechanical system enables the positioning of the cameras' support platform by means of three rotational degrees of freedom and also the control of the vergence angle of each camera. The motors used are stepper motors which are controlled by a system specifically designed and built for this purpose. Each camera's lens has three degrees of freedom , corresponding to zoom, focus and aperture. All the control system electronics as well as the system status data acquisition electronics are integrated in a single control unit. This control unit is connected to a system responsible for the active vision system motion generation.
Lecture Notes in Control and Information Sciences, 2000
In this paper we describe how the combination of several vision sensors can be used to track mult... more In this paper we describe how the combination of several vision sensors can be used to track multiple human targets in a typical surveillance situation. The experimental system devised is made up of a xed camera and a binocular active vision system. The xed camera is used to provide the system with peripheral motion detection capabilities. The human activity is monitored by extracting several parameters that are useful for their classi cation. The system enables the creation of a record based on the type of activity. These logs can be selectively accessed and provide images of the humans in speci c areas.
Lecture Notes in Control and Information Sciences, 1997
This article describes an algorithm for detection of obstacles, through the use of low cost sonar... more This article describes an algorithm for detection of obstacles, through the use of low cost sonar sensors, for the navigation of mobile robots. The use of a grid representation for the cartesian space, associated to probabilistic operations on it, allowed the implementation of a fast, simple and robust algorithm for obstacle detection. This algorithm runs always that the mobile robot navigates, and allows to overcome obstacles that appear unexpectedly in the robot's path. The information used to represent the obstacles is based in the information provided by sonar sensors. A stochastic model for the sonar sensors and their activity, associated to the platform motion, are explored by the algorithm. Some experimental results obtained through the use this algorithm are described at the end of the article.
Tracking in 3D with an active vision system depends on the performance of both motor control and ... more Tracking in 3D with an active vision system depends on the performance of both motor control and vision algorithms. Tracking is performed based on different visual behaviors, namely smooth pursuit and vergence control. A major issue in a system performing tracking is its robustness to partial occlusion of the target as well as its robustness to sudden changes of target trajectory. Another important issue is the reconstruction of the 3D trajectory of the target. These issues can only be dealt with if the performance of the algorithms is evaluated. The evaluation of such performances enable the identification of the limits and weaknesses in the system behavior. In this paper we describe the results of the analysis of a binocular tracking system. To perform the evaluation a control framework was used both for the vision algorithms and for the servo-mechanical system. Due to the geometry changes in an active vision system, the problem of defining and generating system reference inputs has specific features. In this paper we analyze this problem, proposing and justifying a methodology for the definition and generation of such reference inputs. As a result several algorithms were improved and the global performance of the system was also enhanced. This paper proposes a methodology for such an analysis (and resulting enhancements) based on techniques from control theory.
Lecture Notes in Control and Information Sciences, 1998
In this paper we discuss the use of active vision for the development of autonomous systems. Acti... more In this paper we discuss the use of active vision for the development of autonomous systems. Active vision systems are essentially based on biological motivations. Two systems with potential application to surveillance are described. Both systems behave as watchrobots. One of them involves the integration of an active vision system in a mobile platform. The second system can track non-rigid
The robustness of visual behaviors implemented in an active vision sys- tem depends both on the v... more The robustness of visual behaviors implemented in an active vision sys- tem depends both on the vision algorithms and control structure. To improve robustness an evaluation of the performance of the several system algorithms must be done. Performance analysis can be done within the framework of con- trol theory since notions such as stability and controllability can contribute to a
Microprocessors in Robotic and Manufacturing Systems, 1991
Sensor-based robots can increase significantly the scope of applications of robots in manufacturi... more Sensor-based robots can increase significantly the scope of applications of robots in manufacturing namely for flexible assembly. This paper describes the development of a robotic system used for the evaluation of sensory integration techniques and real-time path control strategies. The robotic system has a distributed hierarchical architecture, featuring vision, force, tactile, range and proximity sensors. The approaches for integration of
Object recognition supported by user interaction for service robots
In this paper we describe a real-time gesture recognition algorithm based on contour signature im... more In this paper we describe a real-time gesture recognition algorithm based on contour signature images, generated by collecting a set of contour signatures over time. Each contour signature is generated in each frame by computing the polar coordinates of each point of the contour in relation to the center of mass of that contour. The gesture recognition is reduced to the problem of finding the correspondent contour signature image from a set of previously learned image contour signatures. This is achieved using a Karhunen-Loeve (KL) transform that reduces the set of trained images to a low-dimensional eigenspace, making recognition easier. This method is particularly suitable for the recognition of gestures that can be perfectly described by silhouettes. Since no shape or 3D model is required the proposed method can easily be adapted to recognize several other types of gestures not necessarily performed by humans. The proposed algorithm may be able to extract target gesture motions and classify them at the same time. Preliminary experimental results demonstrate robustness and real-time performance of the proposed algorithm.
Procedings of the British Machine Vision Conference 1993, 1993
A new method for robotics camera calibration based on a linear computation and using a coplanar g... more A new method for robotics camera calibration based on a linear computation and using a coplanar group of calibration points is presented in this paper. This new method suppresses the constraint presented by the RAC calibration model of having an incidence angle between the optical axis and the calibration plane of at least 30 degrees. The parameters are obtained based on a multi-step calculation, and the accuracy of the parameters is strongly increased when used on a iterative calibration. Experimental and simulated analyses results are presented. A comparative analysis between this new method and the RAC method is presented.
In this paper we describe the integration of vision information obtained by a fixed camera and th... more In this paper we describe the integration of vision information obtained by a fixed camera and the information obtained by a binocular active vision system. Both systems acquire images from the same environment. The information redundancy is exploited to achieve a higher degree of robustness in detecting and tracking intruders. Global information about intruders is obtained by the fixed vision system whereas information about a specific intruder (that invades a critical area, for example) is obtained by the binocular system.
Uploads
Papers by Helder Araujo