Downloads & Free Reading Options - Results

This article is from Sensors (Basel, Switzerland) , volume 13 . Abstract In this paper we present a complete spike-based architecture: from a Dynamic Vision Sensor (retina) to a stereo head robotic platform. The aim of this research is to reproduce intended movements performed by humans taking into account as many features as possible from the biological point of view. This paper fills the gap between current spike silicon sensors and robotic actuators by applying a spike processing strategy to the data flows in real time. The architecture is divided into layers: the retina, visual information processing, the trajectory generator layer which uses a neuroinspired algorithm (SVITE) that can be replicated into as many times as DoF the robot has; and finally the actuation layer to supply the spikes to the robot (using PFM). All the layers do their tasks in a spike-processing mode, and they communicate each other through the neuro-inspired AER protocol. The open-loop controller is implemented on FPGA using AER interfaces developed by RTC Lab. Experimental results reveal the viability of this spike-based controller. Two main advantages are: low hardware resources (2% of a Xilinx Spartan 6) and power requirements (3.4 W) to control a robot with a high number of DoF (up to 100 for a Xilinx Spartan 6). It also evidences the suitable use of AER as a communication protocol between processing and actuation.

This article is from Sensors (Basel, Switzerland) , volume 13 . Abstract In this paper we present a complete spike-based architecture: from a Dynamic Vision Sensor (retina) to a stereo head robotic platform. The aim of this research is to reproduce intended movements performed by humans taking into account as many features as possible from the biological point of view. This paper fills the gap between current spike silicon sensors and robotic actuators by applying a spike processing strategy to the data flows in real time. The architecture is divided into layers: the retina, visual information processing, the trajectory generator layer which uses a neuroinspired algorithm (SVITE) that can be replicated into as many times as DoF the robot has; and finally the actuation layer to supply the spikes to the robot (using PFM). All the layers do their tasks in a spike-processing mode, and they communicate each other through the neuro-inspired AER protocol. The open-loop controller is implemented on FPGA using AER interfaces developed by RTC Lab. Experimental results reveal the viability of this spike-based controller. Two main advantages are: low hardware resources (2% of a Xilinx Spartan 6) and power requirements (3.4 W) to control a robot with a high number of DoF (up to 100 for a Xilinx Spartan 6). It also evidences the suitable use of AER as a communication protocol between processing and actuation.

This article is from Sensors (Basel, Switzerland) , volume 13 . Abstract In this paper we present a complete spike-based architecture: from a Dynamic Vision Sensor (retina) to a stereo head robotic platform. The aim of this research is to reproduce intended movements performed by humans taking into account as many features as possible from the biological point of view. This paper fills the gap between current spike silicon sensors and robotic actuators by applying a spike processing strategy to the data flows in real time. The architecture is divided into layers: the retina, visual information processing, the trajectory generator layer which uses a neuroinspired algorithm (SVITE) that can be replicated into as many times as DoF the robot has; and finally the actuation layer to supply the spikes to the robot (using PFM). All the layers do their tasks in a spike-processing mode, and they communicate each other through the neuro-inspired AER protocol. The open-loop controller is implemented on FPGA using AER interfaces developed by RTC Lab. Experimental results reveal the viability of this spike-based controller. Two main advantages are: low hardware resources (2% of a Xilinx Spartan 6) and power requirements (3.4 W) to control a robot with a high number of DoF (up to 100 for a Xilinx Spartan 6). It also evidences the suitable use of AER as a communication protocol between processing and actuation.

Teaching a robot to perceive and navigate in an unstructured natural world is a difficult task. Without learning, navigation systems are short-range and extremely limited. With learning, the robot can be taught to classify terrain at longer distances, but these classifiers can be fragile as well, leading to extremely conservative planning. A robust, high-level learning-based perception system for a mobile robot needs to continually learn and adapt as it explores new environments. To do this, a strong feature representation is necessary that can encode meaningful, discriminative patterns as well as invariance to irrelevant transformations. A simple realtime classifier can then be trained on those features to predict the traversability of the current terrain. One such method for learning a feature representation is discussed in detail in this work. Dimensionality reduction by learning an invariant mapping (DrLIM) is a weakly supervised method for learning a similarity measure over a domain. Given a set of training samples and their pairwise relationships, which can be arbitrarily defined, DrLIM can be used to learn a function that is invariant to complex transformations of the inputs such as shape distortion and rotation. The main contribution of this work is a self-supervised learning process for long-range vision that is able to accurately classify complex terrain, permitting improved strategic planning. As a mobile robot moves through offroad environments, it learns traversability from a stereo obstacle detector. The learning architecture is composed of a static feature extractor, trained offline for a general yet discriminative feature representation, and an adaptive online classifier. This architecture reduces the effect of concept drift by allowing the online classifier to quickly adapt to very few training samples without overtraining. After experiments

Teaching a robot to perceive and navigate in an unstructured natural world is a difficult task. Without learning, navigation systems are short-range and extremely limited. With learning, the robot can be taught to classify terrain at longer distances, but these classifiers can be fragile as well, leading to extremely conservative planning. A robust, high-level learning-based perception system for a mobile robot needs to continually learn and adapt as it explores new environments. To do this, a strong feature representation is necessary that can encode meaningful, discriminative patterns as well as invariance to irrelevant transformations. A simple realtime classifier can then be trained on those features to predict the traversability of the current terrain. One such method for learning a feature representation is discussed in detail in this work. Dimensionality reduction by learning an invariant mapping (DrLIM) is a weakly supervised method for learning a similarity measure over a domain. Given a set of training samples and their pairwise relationships, which can be arbitrarily defined, DrLIM can be used to learn a function that is invariant to complex transformations of the inputs such as shape distortion and rotation. The main contribution of this work is a self-supervised learning process for long-range vision that is able to accurately classify complex terrain, permitting improved strategic planning. As a mobile robot moves through offroad environments, it learns traversability from a stereo obstacle detector. The learning architecture is composed of a static feature extractor, trained offline for a general yet discriminative feature representation, and an adaptive online classifier. This architecture reduces the effect of concept drift by allowing the online classifier to quickly adapt to very few training samples without overtraining. After experiments

Octek 4200 Robot Vision Module, automatic manufacturing and inspection processes - Octek Inc.

Robot Vision Module - Octek, Inc. 1981

Robot Vision Module - Octek, Inc. 1981

6 p. 28 cm

6 p. 28 cm

6 p. 28 cm

6 p. 28 cm

The advent of spacecraft mobile robots-free-flyng sensor platforms and communications devices intended to accompany astronauts or remotely operate on space missions both inside and outside of a spacecraft-has demanded the development of a simple and effective navigation schema. One such system under exploration involves the use of a laser-camera arrangement to predict relative positioning of the mobile robot. By projecting laser beams from the robot, a 3D reference frame can be introduced. Thus, as the robot shifts in position, the position reference frame produced by the laser images is correspondingly altered. Using normalization and camera registration techniques presented in this paper, the relative translation and rotation of the robot in 3D are determined from these reference frame transformations.

The research conducted under this grant has been to investigate techniques for recognizing object that are partially hidden or occluded. It was assumed that geometrical data about the objects was available either as a by- product of the CAD process used to design the object or as the output of a training phase. There were three aspects to this work. First, recognition techniques were developed that used only a single view gray-level image. Second, complementary techniques were developed that used range images. Third, computation times for both of these techniques were studied with the intent of reducing them through the use of parallel processing. The work on gray-level images resulted in a number of new techniques to speedup recognition. These included the concept of saliency and critical point neighborhood, and a fast matching algorithm based on k-d trees. Techniques for recognizing occluded objects that can be at varying distances from the viewer were also developed. The work with range data was aimed at the recognition and pose estimation of three-dimensional industrial parts. The major subproblems addressed were: (1) segmentation and surface parameter extraction, (2) grouping regions into a set (a Convex Region Set), and (3) Convex Region Set matching. The work to develop fast parallel versions of our recognition algorithms made extensive use of a 64 processor NCUBE hypercube multiprocessor. Speedups of 30:1 over VAX 11/780 implementations were obtained for some of the low-level image processing algorithms used in the work.

The research conducted under this grant has been to investigate techniques for recognizing object that are partially hidden or occluded. It was assumed that geometrical data about the objects was available either as a by- product of the CAD process used to design the object or as the output of a training phase. There were three aspects to this work. First, recognition techniques were developed that used only a single view gray-level image. Second, complementary techniques were developed that used range images. Third, computation times for both of these techniques were studied with the intent of reducing them through the use of parallel processing. The work on gray-level images resulted in a number of new techniques to speedup recognition. These included the concept of saliency and critical point neighborhood, and a fast matching algorithm based on k-d trees. Techniques for recognizing occluded objects that can be at varying distances from the viewer were also developed. The work with range data was aimed at the recognition and pose estimation of three-dimensional industrial parts. The major subproblems addressed were: (1) segmentation and surface parameter extraction, (2) grouping regions into a set (a Convex Region Set), and (3) Convex Region Set matching. The work to develop fast parallel versions of our recognition algorithms made extensive use of a 64 processor NCUBE hypercube multiprocessor. Speedups of 30:1 over VAX 11/780 implementations were obtained for some of the low-level image processing algorithms used in the work.

A critical limitation of most of the robots utilized in industrial environments arises due to their inability to utilize sensory feedback. This forces robot operation into totally preprogrammed or teleoperation modes. In order to endow the new generation of robots with higher levels of autonomy techniques for sensing of their work environments and for accurate and efficient analysis of the sensory data must be developed. In this paper detailed development of vision system modules for inspecting various types of meters, both analog and digital, encountered in a robotic inspection and manipulation tasks are described. These modules are tested using industrial robots having multisensory input capability.

conway game ladies turk

Graphics Demo 1 (1987 09 16)(International Robot Crew The Real Vision)

Book Source: Digital Library of India Item 2015.199528 dc.contributor.author: Vasudevan .a dc.date.accessioned: 2015-07-08T13:15:34Z dc.date.available: 2015-07-08T13:15:34Z dc.date.digitalpublicationdate: 2005-08-27 dc.identifier.barcode: 5990010101956 dc.identifier.origpath: /rawdataupload/upload/0101/958 dc.identifier.copyno: 1 dc.identifier.uri: http://www.new.dli.ernet.in/handle/2015/199528 dc.description.scannerno: 14 dc.description.scanningcentre: IIIT, Allahabad dc.description.main: 1 dc.description.tagged: 0 dc.description.totalpages: 70 dc.format.mimetype: application/pdf dc.language.iso: English dc.publisher: Indian Institute Of Technology Kanpur dc.rights: Out_of_copyright dc.source.library: Indian Institute Of Technology Kanpur dc.subject.classification: Technology dc.subject.classification: Electrical Engineering dc.title: Simulation Of A Robot Vision System Based On Two Dimensional Image Analysis

The premarking of parts for subsequent identification by a robot vision system appears to be beneficial as an aid in the automation of certain tasks such as construction in space. A simple, color coded marking system is presented which allows a computer vision system to locate an object, calculate its orientation, and determine its identity. Such a system has the potential to operate accurately, and because the computer shape analysis problem has been simplified, it has the ability to operate in real time.

The scope of the current work is to develop practical sensing implementations for robots operating in complex, partially unstructured environments. A focus in this work is to develop object models and estimation techniques which are specific to requirements of robot locomotion, approach and avoidance, and grasp and manipulation. Such problems have to date received limited attention in either computer or human vision - in essence, asking not only how perception is in general modeled, but also what is the functional purpose of its underlying representations. As in the past, researchers are drawing on ideas from both the psychological and machine vision literature. Of particular interest is the development 3-D shape and motion estimates for complex objects when given only partial and uncertain information and when such information is incrementally accrued over time. Current studies consider the use of surface motion, contour, and texture information, with the longer range goal of developing a fused sensing strategy based on these sources and others.

The scope of the current work is to develop practical sensing implementations for robots operating in complex, partially unstructured environments. A focus in this work is to develop object models and estimation techniques which are specific to requirements of robot locomotion, approach and avoidance, and grasp and manipulation. Such problems have to date received limited attention in either computer or human vision - in essence, asking not only how perception is in general modeled, but also what is the functional purpose of its underlying representations. As in the past, researchers are drawing on ideas from both the psychological and machine vision literature. Of particular interest is the development 3-D shape and motion estimates for complex objects when given only partial and uncertain information and when such information is incrementally accrued over time. Current studies consider the use of surface motion, contour, and texture information, with the longer range goal of developing a fused sensing strategy based on these sources and others.

The scope of the current work is to develop practical sensing implementations for robots operating in complex, partially unstructured environments. A focus in this work is to develop object models and estimation techniques which are specific to requirements of robot locomotion, approach and avoidance, and grasp and manipulation. Such problems have to date received limited attention in either computer or human vision - in essence, asking not only how perception is in general modeled, but also what is the functional purpose of its underlying representations. As in the past, researchers are drawing on ideas from both the psychological and machine vision literature. Of particular interest is the development 3-D shape and motion estimates for complex objects when given only partial and uncertain information and when such information is incrementally accrued over time. Current studies consider the use of surface motion, contour, and texture information, with the longer range goal of developing a fused sensing strategy based on these sources and others.

Final Project for ROB-UY 3203 Robot Vision of NYU submitted by AP6487 & NP2289

Jetson Nano “JetBot” machine learning robot review and demo. Includes hardware, software, Jupyter Lab notebooks for executing Python code, collision detection examples, and some introduction to training a neural network model. The JetBot shown was supplied for review by NVIDIA, but this is not a sponsored video. If you are interested in getting a JetBot, I would strongly recommend starting on its NVIDIA web pages here: https://developer.nvidia.com/embedded/jetson-nano-developer-kit The JetBot Wiki, which contains all of the information required to build, setup and run vision recognition and machine learning on a JetBot, is here: https://github.com/NVIDIA-AI-IOT/jetbot/wiki There is also info on full JetBot kits here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/jetbot-ai-robot-kit/ My previous review of the NVIDIA Jetson Nano SBC is here: https://www.youtube.com/watch?v=tgGELfqigsc And my “Jetson Nano: Vision Recognition Neural Network Demo” video is here: https://www.youtube.com/watch?v=k5pXXmTkPNM I also have an introduction to AI video here: https://www.youtube.com/watch?v=W2ey_4_DHuc More videos on single board computers and broader computing topics can be found on the ExplainingComputers channel: https://www.youtube.com/user/explainingcomputers/videos You may also like my other channel, ExplainingTheFuture, at: https://www.youtube.com/user/explainingthefuture/videos #JetBot #JetsonNano #NVIDIA #ExplainingComputers

A positioning and navigation method based on stereo vision, compass and encoders is presented. It has been used in an autonomous mobile robot which developed by the group of the authors. The compass and encoders are used complementarily each other to get correct position. 3D vision-based path-planning makes the robot walk along a better path each calculation cycle and avoid bumping other objects. This navigation system is cheap and convenient to create, and is effective in practice. Complex positioning, environment mapping, path-planning and behavior-based navigation algorithm are discussed is this paper.

Book Source: Digital Library of India Item 2015.193183 dc.contributor.author: Musunur Laxmi Prasad dc.date.accessioned: 2015-07-08T02:18:07Z dc.date.available: 2015-07-08T02:18:07Z dc.date.digitalpublicationdate: 2005-08-21 dc.identifier.barcode: 1990010086716 dc.identifier.origpath: /rawdataupload/upload/0086/716 dc.identifier.copyno: 1 dc.identifier.uri: http://www.new.dli.ernet.in/handle/2015/193183 dc.description.scannerno: 14 dc.description.scanningcentre: IIIT, Allahabad dc.description.main: 1 dc.description.tagged: 0 dc.description.totalpages: 88 dc.format.mimetype: application/pdf dc.language.iso: English dc.publisher: Indian Institute Of Technology Kanpur dc.rights: Out_of_copyright dc.source.library: Indian Institute Of Technology Kanpur dc.subject.classification: Technology dc.subject.classification: Engineering. Technology In General dc.subject.classification: Mechanical Engineering In General. Nuclear Technology. Electrical Engineering. Machinery dc.title: Eye Kanpur -grippiing Polyhedral Object Though Robot Vision

A prototype twin-camera stereo vision system for autonomous robots has been developed at Goddard Space Flight Center. Standard charge coupled device (CCD) imagers are interfaced with commercial frame buffers and direct memory access to a computer. The overlapping portions of the images are analyzed using photogrammetric techniques to obtain information about the position and orientation of objects in the scene. The camera head consists of two 510 x 492 x 8-bit CCD cameras mounted on individually adjustable mounts. The 16 mm efl lenses are designed for minimum geometric distortion. The cameras can be rotated in the pitch, roll, and yaw (pan angle) directions with respect to their optical axes. Calibration routines have been developed which automatically determine the lens focal lengths and pan angle between the two cameras. The calibration utilizes observations of a calibration structure with known geometry. Test results show the precision attainable is plus or minus 0.8 mm in range at 2 m distance using a camera separation of 171 mm. To demonstrate a task needed on Space Station Freedom, a target structure with a movable I beam was built. The camera head can autonomously direct actuators to dock the I-beam to another one so that they could be bolted together.

The role of unmanned aerial vehicles (UAVs) in military, commercial and recreational applications is continuously evolving as developments in technology increase capabilities. The research herein presents an inexpensive computer-vision-based solution for detection and classification of a stationary target with a mobile aerial sensor as a prototyping platform. The main goal of this system is to use commercial-off-the-shelf and open-source components to reduce design complexity to provide a legacy product for future development of specific capabilities. Color imagery collected during flight using a low-resolution camera is used to test the application of a simple algorithm against a commercially available and low cost sensor. Original image processing algorithms that leverage the existing body of works in the open-source community are developed and tested within the Systems Engineering construct. System architecture leverages a modular approach that can be easily modified and adapted to changing requirements and objectives. Conclusions are drawn and recommendations for further study and system development are presented.

this is a Stereo vision set up for the Baxter robot. For more information on installation go to http://wiki.lofarolabs.com/index.php/Baxter into the build section the point of this is that you see what the robot sees but at his scale. The scale was figured out by the difference in the length of our arms because Baxters arms are long it means his eyes are 11 cm apart.

robot vision

Introduction Nowadays, machine vision systems are extensively used in agriculture. The application of this technology in the field can help preserve agricultural resources while reducing manual labor and production costs. In the field of agricultural automation, accurately detecting crop rows is recognized as a crucial and challenging issue for weed identification and the automatic guidance of machines. Therefore, it is necessary to explore practical solutions to optimize this process. Hence, the purpose of this study is the precise identification of basil cultivation rows to enable the automatic navigation of robots in the cultivation field. Materials and Methods In the first stage of this research, six images from each growth period of basil plants (third, fourth, and fifth week) were taken and weeds were removed from the area between the crop rows using three different methods of area opening, dimensional removal, and masking. In the next stage, six images of crop rows without weeds were examined by performing image processing operations and implementing several routing algorithms, namely, Hough transform, wavelet transform, Gabor filter, linear regression, and an additional algorithm proposed in this study. The output of each of these algorithms was compared with the ideal path identified by the user. For this purpose, after capturing an image, green areas were extracted from it by performing the segmentation process. By applying each of the routing algorithms to the image, plant cultivation lines were identified and their equations were determined. Finally, the performance of the designed robot was evaluated using the most appropriate routing algorithm. Results and Discussion Examining the performance of three different methods of weed removal in three periods of plant growth (third, fourth, and fifth week) showed that during this interval, the masking method had the lowest error rate compared to the ideal path and the shortest average operation time of 1.64 seconds, followed by the dimensional removal and the area opening methods. Comparing the routes detected by different routing algorithms with the ideal routes and according to the results of the t-test at 5% probability level, the order of the studied routing methods from the most superior is as follows: the proposed algorithm, Gabor filter, linear regression, Hough transform and wavelet transform algorithm. Overall, the proposed algorithm had the highest rate of adaptation to the ideal path (with an average error of 3.65 pixels) and the shortest operation time (4.79 seconds) and was selected as the most appropriate routing algorithm and the performance of the designed robot was evaluated using it. Conclusion A reliable crop row detection algorithm can reduce production costs and preserve the environment. In this study, the masking method was used for removing weeds from the images. The new proposed routing algorithm has superior performance when compared with common routing algorithms such as the Gabor filter, linear regression, Hough transform, and wavelet transform. Additionally, it was shown that the designed robot using the proposed algorithm (with an average error of 3.65 pixels) has the desired performance. Acknowledgment The authors express appreciation for the financial support provided by Tarbiat Modares University.

A simple approach to robot applications software is described. The idea is to use commercially available software and hardware wherever possible to minimize system costs, schedules and risks. The U.S. has been slow in the adaptation of robots and flexible automation compared to the fluorishing growth of robot implementation in Japan. The U.S. can benefit from this approach because of a more flexible array of vision guided robot technologies.

A simple approach to robot applications software is described. The idea is to use commercially available software and hardware wherever possible to minimize system costs, schedules and risks. The U.S. has been slow in the adaptation of robots and flexible automation compared to the fluorishing growth of robot implementation in Japan. The U.S. can benefit from this approach because of a more flexible array of vision guided robot technologies.

The concept and design of a dynamic test platform for development and evluation of a robot vision system is discussed. The platform is to serve as a diagnostic and developmental tool for future work with the RPI Mars Rover's multi laser/multi detector vision system. The platform allows testing of the vision system while its attitude is varied, statically or periodically. The vision system is mounted on the test platform. It can then be subjected to a wide variety of simulated can thus be examined in a controlled, quantitative fashion. Defining and modeling Rover motions and designing the platform to emulate these motions are also discussed. Individual aspects of the design process are treated separately, as structural, driving linkages, and motors and transmissions.

El proyecto busca desarrollar un problema concerniente a robots cosechadores que pueda ser resuelto utilizando visión computacional, seleccionando un producto en especifico que sea de la región y sea de fácil acceso, eligiendo resolver lo relacionado a la detección de  jitomates maduros en un invernadero. Se realiza una investigación a modo del estado del arte para conocer los principales exponentes de los robots cosechadores y ver soluciones posibles al problema, así como una investigación documental de las soluciones posibles y enfoques usados para resolver un problema de segmentación y detección, con soluciones tanto con  procesamiento de imagen como con el uso de redes neuronales. En el presente documento se desarrollan diversas técnicas novedosas utilizadas en redes neuronales convolucionales, presentando un modelo mejorado que tiene ventajas sobre el original en cuestión de entrenamiento.

Response to the Devlelop 3D Make A 'Bot' 2013 robot design competition.

We address the problem of autonomously learning controllers for vision-capable mobile robots. We extend McCallum's (1995) Nearest-Sequence Memory algorithm to allow for general metrics over state-action trajectories. We demonstrate the feasibility of our approach by successfully running our algorithm on a real mobile robot. The algorithm is novel and unique in that it (a) explores the environment and learns directly on a mobile robot without using a hand-made computer model as an intermediate step, (b) does not require manual discretization of the sensor input space, (c) works in piecewise continuous perceptual spaces, and (d) copes with partial observability. Together this allows learning from much less experience compared to previous methods.

The Semantic Robot Vision Competition provided an excellent opportunity for our research lab to integrate our many ideas under one umbrella, inspiring both collaboration and new research. The task, visual search for an unknown object, is relevant to both the vision and robotics communities. Moreover, since the interplay of robotics and vision is sometimes ignored, the competition provides a venue to integrate two communities. In this paper, we outline a number of modifications to the competition to both improve the state-of-the-art and increase participation.

This article is from Sensors (Basel, Switzerland) , volume 11 . Abstract This study aimed to develop a novel 3D parallel mechanism robot driven by three vertical-axial pneumatic actuators with a stereo vision system for path tracking control. The mechanical system and the control system are the primary novel parts for developing a 3D parallel mechanism robot. In the mechanical system, a 3D parallel mechanism robot contains three serial chains, a fixed base, a movable platform and a pneumatic servo system. The parallel mechanism are designed and analyzed first for realizing a 3D motion in the X-Y-Z coordinate system of the robot’s end-effector. The inverse kinematics and the forward kinematics of the parallel mechanism robot are investigated by using the Denavit-Hartenberg notation (D-H notation) coordinate system. The pneumatic actuators in the three vertical motion axes are modeled. In the control system, the Fourier series-based adaptive sliding-mode controller with H∞ tracking performance is used to design the path tracking controllers of the three vertical servo pneumatic actuators for realizing 3D path tracking control of the end-effector. Three optical linear scales are used to measure the position of the three pneumatic actuators. The 3D position of the end-effector is then calculated from the measuring position of the three pneumatic actuators by means of the kinematics. However, the calculated 3D position of the end-effector cannot consider the manufacturing and assembly tolerance of the joints and the parallel mechanism so that errors between the actual position and the calculated 3D position of the end-effector exist. In order to improve this situation, sensor collaboration is developed in this paper. A stereo vision system is used to collaborate with the three position sensors of the pneumatic actuators. The stereo vision system combining two CCD serves to measure the actual 3D position of the end-effector and calibrate the error between the actual and the calculated 3D position of the end-effector. Furthermore, to verify the feasibility of the proposed parallel mechanism robot driven by three vertical pneumatic servo actuators, a full-scale test rig of the proposed parallel mechanism pneumatic robot is set up. Thus, simulations and experiments for different complex 3D motion profiles of the robot end-effector can be successfully achieved. The desired, the actual and the calculated 3D position of the end-effector can be compared in the complex 3D motion control.

SciPy 2010 Day 1 Main Track 5 A High Performance Robot Vision Algorithm Implemented in Python S. Chris Colbert

This paper documents our progress during the first year of work on our original proposal entitled 'A Scalable Distributed Approach to Mobile Robot Vision'. We are pursuing a strategy for real-time visual identification and tracking of complex objects which does not rely on specialized image-processing hardware. In this system perceptual schemas represent objects as a graph of primitive features. Distributed software agents identify and track these features, using variable-geometry image subwindows of limited size. Active control of imaging parameters and selective processing makes simultaneous real-time tracking of many primitive features tractable. Perceptual schemas operate independently from the tracking of primitive features, so that real-time tracking of a set of image features is not hurt by latency in recognition of the object that those features make up. The architecture allows semantically significant features to be tracked with limited expenditure of computational resources, and allows the visual computation to be distributed across a network of processors. Early experiments are described which demonstrate the usefulness of this formulation, followed by a brief overview of our more recent progress (after the first year).

Integration of skills into an autonomous robot that performs a complex task is described. Time constraints prevented complete integration of all the described skills. The biggest problem was tuning the sensor-based region-finding algorithm to the environment involved. Since localization depended on matching regions found with the a priori map, the robot became lost very quickly. If the low level sensing of the world is not working, then high level reasoning or map making will be unsuccessful.

Integration of skills into an autonomous robot that performs a complex task is described. Time constraints prevented complete integration of all the described skills. The biggest problem was tuning the sensor-based region-finding algorithm to the environment involved. Since localization depended on matching regions found with the a priori map, the robot became lost very quickly. If the low level sensing of the world is not working, then high level reasoning or map making will be unsuccessful.

This paper presents an algorithm to find the correspondences of points representing dominant feature in robot stereo vision. The algorithm consists of two main steps: dominant point extraction and dominant point matching. In the feature extraction phase, the algorithm utilizes the widely used Moravec Interest Operator and two other operators: the Prewitt Operator and a new operator called Gradient Angle Variance Operator. The Interest Operator in the Moravec algorithm was used to exclude featureless areas and simple edges which are oriented in the vertical, horizontal, and two diagonals. It was incorrectly detecting points on edges which are not on the four main directions (vertical, horizontal, and two diagonals). The new algorithm uses the Prewitt operator to exclude featureless areas, so that the Interest Operator is applied only on the edges to exclude simple edges and to leave interesting points. This modification speeds-up the extraction process by approximately 5 times. The Gradient Angle Variance (GAV), an operator which calculates the variance of the gradient angle in a window around the point under concern, is then applied on the interesting points to exclude the redundant ones and leave the actual dominant ones. The matching phase is performed after the extraction of the dominant points in both stereo images. The matching starts with dominant points in the left image and does a local search, looking for corresponding dominant points in the right image. The search is geometrically constrained the epipolar line of the parallel-axes stereo geometry and the maximum disparity of the application environment. If one dominant point in the right image lies in the search areas, then it is the corresponding point of the reference dominant point in the left image. A parameter provided by the GAV is thresholded and used as a rough similarity measure to select the corresponding dominant point if there is more than one point the search area. The correlation is used as a final decision tool when there is still more than one point in the search area. If there is no dominant point in the search area of if the points in the search area are below a correlation threshold, then the dominant point in the reference image is occluded and can not be corresponded. The algorithm has been modeled, implemented and shown to be fast, robust and parallel. The parallelism is created from three main features: locality of the operators; a memory optimization scheme; and the ability to fully parallelize the extraction phase which is the most computational intensive task in the algorithm. The last feature is achieved by performing the extraction phase on the two images simultaneously.

This paper presents an algorithm to find the correspondences of points representing dominant feature in robot stereo vision. The algorithm consists of two main steps: dominant point extraction and dominant point matching. In the feature extraction phase, the algorithm utilizes the widely used Moravec Interest Operator and two other operators: the Prewitt Operator and a new operator called Gradient Angle Variance Operator. The Interest Operator in the Moravec algorithm was used to exclude featureless areas and simple edges which are oriented in the vertical, horizontal, and two diagonals. It was incorrectly detecting points on edges which are not on the four main directions (vertical, horizontal, and two diagonals). The new algorithm uses the Prewitt operator to exclude featureless areas, so that the Interest Operator is applied only on the edges to exclude simple edges and to leave interesting points. This modification speeds-up the extraction process by approximately 5 times. The Gradient Angle Variance (GAV), an operator which calculates the variance of the gradient angle in a window around the point under concern, is then applied on the interesting points to exclude the redundant ones and leave the actual dominant ones. The matching phase is performed after the extraction of the dominant points in both stereo images. The matching starts with dominant points in the left image and does a local search, looking for corresponding dominant points in the right image. The search is geometrically constrained the epipolar line of the parallel-axes stereo geometry and the maximum disparity of the application environment. If one dominant point in the right image lies in the search areas, then it is the corresponding point of the reference dominant point in the left image. A parameter provided by the GAV is thresholded and used as a rough similarity measure to select the corresponding dominant point if there is more than one point the search area. The correlation is used as a final decision tool when there is still more than one point in the search area. If there is no dominant point in the search area of if the points in the search area are below a correlation threshold, then the dominant point in the reference image is occluded and can not be corresponded. The algorithm has been modeled, implemented and shown to be fast, robust and parallel. The parallelism is created from three main features: locality of the operators; a memory optimization scheme; and the ability to fully parallelize the extraction phase which is the most computational intensive task in the algorithm. The last feature is achieved by performing the extraction phase on the two images simultaneously.

The role of unmanned aerial vehicles (UAVs) in military, commercial and recreational applications is continuously evolving as devel-opments in technology increase capabilities. The research herein presents an inexpensive computer-vision-based solution for detection and classification of a stationary target with a mobile aerial sensor as a prototyping platform. The main goal of this system is to use commercial-off-the-shelf and open-source components to reduce design complexity to provide a legacy product for future develop-ment of specific capabilities. Color imagery collected during flight using a low-resolution camera is used to test the application of a simple algorithm against a commercially available and low cost sensor. Original image processing algorithms that leverage the existing body of works in the open-source community are developed and tested within the Systems Engineering construct. System architec-ture leverages a modular approach that can be easily modified and adapted to changing requirements and objectives. Conclusions are drawn and recommendations for further study and system development are presented.

The JPL Robot Vision Library (JPLV) provides real-time robot vision algorithms for developers who are not vision specialists. The package includes algorithms for stereo ranging, visual odometry and unsurveyed camera calibration, and has unique support for very wideangle lenses

This paper presents a novel robot vision architecture for perceiving generic 3D clothes configurations. Our architecture is hierarchically structured, starting from low-level curvatures, across mid-level geometric shapes \& topology descriptions; and finally approaching high-level semantic surface structure descriptions. We demonstrate our robot vision architecture in a customised dual-arm industrial robot with our self-designed, off-the-self stereo vision system, carrying out autonomous grasping and dual-arm flattening. It is worth noting that the proposed dual-arm flattening approach is unique among the state-of-the-art robot autonomous system, which is the major contribution of this paper. The experimental results show that the proposed dual-arm flattening using stereo vision system remarkably outperforms the single-arm flattening and widely-cited Kinect-based sensing system for dexterous manipulation tasks. In addition, the proposed grasping approach achieves satisfactory performance on grasping various kind of garments, verifying the capability of proposed visual perception architecture to be adapted to more than one clothing manipulation tasks.