Object Location in Closed Environments for Robots Using an Iconographic Base

This paper presents an innovative visual odometry system for fast motion estimation of autonomous space exploration vehicles. In order to improve a vehicle position recovery, the use of accelerometers and gyros is usually supported by wheels odometry. The visual system can improve the reliability of the overall rover navigation system when used in conjunction with a classical odometry system, usually affected by drift or sliding problems. The visual odometry system is based on artificial vision and reconstructs the rover position through the analysis of sequences of images acquired by two stereo cameras. The proposed algorithm is based on stereo geometry and discrete optical ow computation. From stereo geometry it is possible to locate the image points in the three dimensional environment, while from the optical ow the relative motion of the camera with respect to these points is computed. Moreover, in order to reduce the computational burden usually associated with the optical ow computation, it has been decided to first search the images for features and then to compute the discrete optical ow only associated to them. The features detection and tracking is done by using the KLT algorithm. Tests made with low resolution cameras, showing encouraging results, are presented.

Programming of industrial robots is normally carried out via so-called Teach or Offline programming methodologies. Both methodologies have associated weaknesses like a high time consumption for programming on sculptured surfaces (especially Teach), and both methodologies cannot, at the same time, carry information of the real work piece and the ideal work piece (CAD-model). To be able to see both the real and ideal world is especially important when the industrial robot is used in manufacturing processes like machining, grinding, deburring etc. where the difference between the real and ideal world represents the material removal. This paper presents a new vision based programming methodology which combines information of the real and ideal world, especially adapted for robot grinding and deburring operations. Further, the presented methodology is especially developed with simplicity in mind when it comes to man-machine communication. Thus, a standard marker pen can be used by the operator to draw the robot path directly onto the work piece. This line is captured by a vision system and finally leads to the generation of the robot path. The presented methodology is a further development of the previous work of the authors as presented (Solvang et al., 2007 and Sziebig, 2007), especially concerning optimization of the robot tool positioning with respect to a work piece surface.

Limited fossil energy resources and the potential danger of nuclear power plants led to growing popularity of solar energy. In Switzerland, Building Integrated Photovoltaic (BIPV) is expected to be responsible for up to quarter of the energy production from renewable resources by the time of 2035. In order to protect the existing natural landscape, BIPV must be concentrated in urban spaces, which means that certain amount of existing building envelopes have to be turned into energy generators. There is a growing concern about BIPV retrofits because they may change the visual appearance of the existing city images to a large extend and/or in a negative way. In order to manage the potential visual impact resulting from BIPV expansion in urban spaces, evaluation methods should be able to measure it appropriately. Existing evaluation methods show insufficiencies for this purpose: either they cannot guarantee objectivity and continuity of evaluation standards throughout assessments of different BIPV projects, because their qualitative criteria are vulnerable to subjective preferences; or they only use formal design parameters to evaluate the visual integration quality of BIPV and therefore lack neuroscientific base. In order to tackle these insufficiencies, an objective evaluation method is proposed that is capable of measuring the BIPV visual impact in building retrofits in a quantified approach based on neuroscience knowledge. The assessment should be made in concept phase of the project, so as to identify the BIPV designs that have the least negative visual impact. The proposed evaluation method integrates saliency model, which imitates the mechanism of human visual attention, into assessment procedures. First, the probability of a BIPV installation attracting human visual attention in the respective visual scene is calculated quantitatively with the saliency model. Then the modifications of saliency values in this very visual scene before and after the BIPV retrofit are assessed. In the end, the modifications of saliency values are transformed into BIPV visual impact and objectively expressed as single values. The analyses are based on renderings generated from RADIANCE and programming in MATLAB. This method is demonstrated with a small case study that simultaneously serves as proof-of-concept. The proposed evaluation method is applied on a realistic case study: BIPV designs for a church roof. In total, 5 designs were developed with variations in BIPV installation location, roof coverage percentage, module size, PV material und design approach. The lowest visual impact value was induced by the BIPV design with the most careful and considerate integration approach, the design with the boldest integration approach obtained the highest visual impact value. The evaluation method proved to be feasible to a large extend. It is believed that the synergy between architecture and neuroscience can contribute to a growing understanding of human responses to the built environment. Hopefully the findings from this thesis can help in minimizing the negative visual impact induced by BIPV expansion in urban spaces, and also aid architects in gaining new understandings for visual aspects in architecture design.

Near-infrared (NIR) synaptic devices integrate NIR optical sensitivity and synaptic plasticity, emulating the basic biomimetic function of the human visual system and showing great potential in NIR artificial vision systems. However, the lack of semiconductor materials with appropriate band gaps for NIR photodetection and effective strategies for fabricating devices with synaptic behaviors limit the further development of NIR synaptic devices. Here, a two-terminal NIR synaptic device consisting of the In2Se3/MoS2 heterojunction has been constructed, and it exhibits fundamental synaptic functions. The reduced band gap and potential barrier of In2Se3/MoS2 heterojunctions are essential for NIR synaptic plasticity. In addition, the NIR synaptic properties of In2Se3/MoS2 heterojunctions under strain have been studied systematically. The ΔEPSC of the In2Se3/MoS2 synaptic device can be improved from 38.4% under no strain to 49.0% under a 0.54% strain with a 1060 nm illumination for 1 s at 100 mV. Furthermore, the artificial NIR vision system consisting of a 10 × 10 In2Se3/MoS2 device array has been fabricated, exhibiting image sensing, learning, and storage functions under NIR illumination. This research provides new ideas for the design of flexible NIR synaptic devices based on 2D materials and presents many opportunities in artificial intelligence and NIR vision systems.

This paper presents a test environment enabling the study of factors affecting on the success of a robotic precision assembly work cycle. The developed testing environment measures forces and torques occurring during the assembly, and uses a system based on machine vision to measure the repeatability of work piece positioning. The testing environment is capable of producing exactly known artificial positioning errors in four degrees-of-freedom to simulate errors in work-piece positioning accuracy. The testing environment also measures the total duration of the robot work cycle as well as the durations of all essential phases of the work cycle. The testing environment is best suited for light assembly operations and has measurement ranges of ±36 N and ±0.5 Nm and the vision system has a field-of-view mm.The latter part of this paper presents the results of the research done in order to find out how some selected factors affect the assembly forces of robotic assembly. These factors include work piece and process parameters such as work piece material and design (chamfered/straight), positioning tolerances, and robot insertion motion speed.Key wordsAssembly forceassembly process testingwork cycle timepositioning errors

In vision system of laser remanufacturing robot, binocular stereovision system combining structured light technique is used, Which applies to measure Laser Remanufacturing workpieces. The vision system which includes two cameras, a laser balopticon and a computer is fitted on the bottom of the robot, and its position is controlled by robot teach pedant. The image collected by cameras is transferred to a computer. Calibration and matching is the key to the work of the visual system. Calibration, including camera calibration and robot hand-eye calibration, is completed by MATLAB programming. The method of combining structured light with binocular stereo vision system is studied: let line structured light scanning the surface of the work piece, collect image, extract sub-pixel center of structured light by gravity algorithm and complete the image matching by epipolar geometry constraint relations. The process can be completed by MATLAB. The results of the 3D measurements of the work piece show that the system can make the robot recognize the defects and the shape of the work piece automatically, so that the vision system can satisfy the requirements auto-detection remanufacturing.

Automatic Identification of Needle Insertion Site in Epidural Anesthesia with a Cascading Classifier

Backpropagation neural networks have been developed for detection of geological lineaments in the Landsat Thematic Mapper (TM) imagery of the Canadian Shield using edge images as input and digitized lineament maps as the desired output. Lineament detection is a challenging problem for traditional image processing and pattern recognition techniques. Many linear features observable in geological image data do not represent lineaments, and the presence and extent of lineaments must be inferred from contextual information. In order to compare the ability of neural networks and conventional classifiers to recognize lineaments prior to performing edge/line element grouping operations, various gradient and curvature features are extracted from the image data set. Selected features from this group formed the inputs to backpropagation neural networks, linear discriminant classifiers, and nearest-neighbor classifiers. The neural network results were compared with the results obtained using conventional classifiers for sample training and test sets. The trained neural network was then applied to the edge image to mask out those edge points which had been classified as non- lineament points.© (1991) COPYRIGHT SPIE--The International Society for Optical Engineering. Downloading of the abstract is permitted for personal use only.

Identification of critical samples of stochastic processes towards feasible structural reliability applications

Pattern recognition of nucleated cells from the peripheral blood

Automated detection of exudates and macula for grading of diabetic macular edema

The goal of this project was to detect the intensity of traffic on a road at different times of the day during daytime. Although the work presented utilized images from a section of a highway, the results of this project are intended for making decisions on the type of intervention necessary on any given road at different times for traffic control, such as installation of traffic signals, duration of red, green and yellow lights at intersections, and assignment of traffic control officers near school zones or other relevant locations. In this project, directional patterns are used to detect and count the number of cars in traffic images over a fixed area of the road to determine local traffic intensity. Directional patterns are chosen because they are simple and common to almost all moving vehicles. Perspective vision effects specific to each camera orientation has to be considered, as they affect the size and direction of patterns to be recognized. In this work, a simple and fast algorithm has been developed based on horizontal directional pattern matching and perspective vision adjustment. The results of the algorithm under various conditions are presented and compared in this paper. Using the developed algorithm, the traffic intensity can accurately be determined on clear days with average sized cars. The accuracy is reduced on rainy days when the camera lens contains raindrops, when there are very long vehicles, such as trucks or tankers, in the view, and when there is very low light around dusk or dawn.

Guiding and assisting the visually impaired has been a significant problem. Yet, this problem has received insufficient attention. With the pervasion of smartphones, we attempt to utilize their powerful abilities to assist the visually impaired. Existing applications for assisting the visually impaired are mostly obstacle detection, most of which rely on image processing or pattern recognition techniques. However, these techniques are either too slow to achieve real-time processing or impractical due to heavy computational and storage overheads. In this paper, we first propose an acoustic sound based obstacle echolocation ranging method using chirp signals. Next, we present a crowdsourcing framework for obstacle detection, where the surrounding smartphones are utilized as an array of sensors for better results. Combining this framework with a modified TDOA algorithm and Least-Squares method, we thus propose a LsTDOA scheme to effectively locate 3-D coordinates of the obstacle. Furthermore, we take the priori knowledge of possible errors into consideration, and then propose weighted-LsTDOA(WLsTDOA) to improve the accuracy of obstacle estimation. We perform simulation experiments to demonstrate the accuracy and reliability of our scheme, which can be achieved in real-time.

A supercomputer-based software which enables reconstruction of a three-dimensional shape without human assistance using a stereo-pair of scanning electron micrographs has been developed. It is based upon computer image processing and pattern recognition techniques, and can reconstruct a three-dimensional shape of a wide variety of fracture surfaces of corrosion fatigue as well as crack initiation sites in an aggressive environment line a corrosion pit. We discuss the optimum algorithm according to the morphology of the fracture surfaces. A sequential similarity detection algorithm is available in transgranular fatigued surfaces, whereas in the case of corrosion fatigue fracture surfaces and crack initiation sites, the mutual correlation coefficient technique is well suited. The shape and size of the window area, and the run time required in reconstruction are also discussed.

Autonomous navigation has been a difficult problem for traditional vision and robotic techniques, primarily because of the noise and variability associated with real world scenes. Autonomous navigation systems based on traditional image processing and pattern recognition techniques often perform well under certain conditions but have problems with others. Part of the difficulty stems from the fact that the processing performed by these systems remains fixed across various driving situations.KeywordsHide UnitAutonomous NavigationDriving SituationRoad EdgeFixed WidthThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.