SEARCH · Engineering Papers
Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.
Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.
The goal of this research project is the development of a computer vision system to monitor and control life sciences experimentation on board space stations. The vision system is organized as a multiprocessor system with distributed processes selectively analyzing hierarchical imagery in order to monitor and control the appropriate instrumentation.
For automation and robotics in the Space Station era, NASA's Johnson Space Center is pursuing several means of synthetic vision. The optical correlator is one such means. The deformable mirror device will form the basis of the first correlator in this project. In-house and contracted effort is being used. Initial in-house activities will concentrate on an impulse deconvolution technique and on a programmable retina for spatial remappings of an image prior to correlation. The retina will permit a form of edge extraction and other primitive operations. Additionally, it will be used as a research tool for assisting persons with low vision.
Compact, lightweight video camera head constructed by remounting lens and charge-coupled-device image detector from small commercial video camera in separate assembly. Useful in robotics, artificial vision, and vision guidance systems. Designed to be mounted on visor of helmet to monitor motions of eyes in experiments on vestibulo-ocular reflexes.
Image data from two distinct windows used to locate weld features. Analyzer part of vision system described in companion article, "Image Control in Automatic Welding Vision System" (MFS-26035). Horizontal video lines define windows for viewing unwelded joint and weld pool. Data from picture elements outside windows not processed. Widely-separated local features carry no significance, but closely spaced features indicate welding feature. Image processor assigns confidence level to group of local features according to spacing and pattern.
With a goal of providing means for accelerating the image processing, machine vision, and testing of human vision models, an image transform was designed, which makes it possible to map an image into a set of images that vary in resolution and orientation. Each pixel in the output may be regarded as the simulated response of a neuron in human visual cortex. The transform is amenable to a number of shortcuts that greatly reduce the amount of computation.
Hybrid optical correlators, which use spatial light modulators and solid state imagers to interface computers and optical Fourier correlators, will allow the rapid changing of filters and digital pre- or post-correlation processing in vision systems. However, optical correlation (particulary with phase-only filtration) is notoriously sensitive to mismatches of rotation and magnification between the reference image and the input image. Such sharpness is not always desirable if the correlation pattern is to be used in a control system. Two methods of conforming the response of an inherently sharp-response optical correlation to an in plane rotation between reference and object images are discussed. The objective is to condition the correlation for better suitability to a vision-controlled system. The first method is controlled convolutional blurring of a reference image followed by construction of a matched filter (perhaps implemented as a phase-only filter). An alternative is discussed in more detail: it is a synthesis from sharp filters, so weighted as to produce a specified off-center behavior.
The use of robotic manipulators for the mating and demating of umbilical fuel lines to the Space Shuttle Vehicle prior to launch is investigated. Force feedback control is necessary to minimize the contact forces which develop during mating. The objective is to develop and demonstrate a working robotic force control system. Initial experimental force control tests with an ASEA IRB-90 industrial robot using the system's Adaptive Control capabilities indicated that control stability would by a primary problem. An investigation of the ASEA system showed a 0.280 second software delay between force input commands and the output of command voltages to the servo system. This computational delay was identified as the primary cause of the instability. Tests on a second path into the ASEA's control computer using the MicroVax II supervisory computer show that time delay would be comparable, offering no stability improvement. An alternative approach was developed where the digital control system of the robot was disconnected and an analog electronic force controller was used to control the robot's servosystem directly, allowing the robot to use force feedback control while in rigid contact with a moving three-degree-of-freedom target. An alternative approach was developed where the digital control system of the robot was disconnected and an analog electronic force controller was used to control the robot's servo system directly. This method allowed the robot to use force feedback control while in rigid contact with moving three degree-of-freedom target. Tests on this approach indicated adequate force feedback control even under worst case conditions. A strategy to digitally-controlled vision system was developed. This requires switching between the digital controller when using vision control and the analog controller when using force control, depending on whether or not the mating plates are in contact.
The objective of preprocessing for machine vision is to extract intrinsic target properties. The most important properties ordinarily are structure and reflectance. Illumination in space, however, is a significant problem as the extreme range of light intensity, stretching from deep shadow to highly reflective surfaces in direct sunlight, impairs the effectiveness of standard approaches to machine vision. To overcome this critical constraint, an image coding scheme is being investigated which combines local intensity adaptivity, image enhancement, and data compression. It is very effective under the highly variant illumination that can exist within a single frame or field of view, and it is very robust to noise at low illuminations. Some of the theory and salient features of the coding scheme are reviewed. Its performance is characterized in a simulated space application, the research and development activities are described.
A robot system called RAMBO is being developed. It is equipped with a camera, which, given a sequence of simple tasks, can perform these tasks on a moving object. RAMBO is given a complete geometric model of the object. A low level vision module extracts and groups characteristic features in images of the object. The positions of the object are determined in a sequence of images, and a motion estimate of the object is obtained. This motion estimate is used to plan trajectories of the robot tool to relative locations nearby the object sufficient for achieving the tasks. More specifically, low level vision uses parallel algorithms for image enchancement by symmetric nearest neighbor filtering, edge detection by local gradient operators, and corner extraction by sector filtering. The object pose estimation is a Hough transform method accumulating position hypotheses obtained by matching triples of image features (corners) to triples of model features. To maximize computing speed, the estimate of the position in space of a triple of features is obtained by decomposing its perspective view into a product of rotations and a scaled orthographic projection. This allows the use of 2-D lookup tables at each stage of the decomposition. The position hypotheses for each possible match of model feature triples and image feature triples are calculated in parallel. Trajectory planning combines heuristic and dynamic programming techniques. Then trajectories are created using parametric cubic splines between initial and goal trajectories. All the parallel algorithms run on a Connection Machine CM-2 with 16K processors.
Recent developments in the theory of early vision are described which lead from the formulation of the motion problem as an ill-posed one to its solution by minimizing certain 'cost' functions. These cost or energy functions can be mapped onto simple analog and digital resistive networks. It is shown how the optical flow can be computed by injecting currents into resistive networks and recording the resulting stationary voltage distribution at each node. These networks can be implemented in cMOS VLSI circuits and represent plausible candidates for biological vision systems.
The present computations for three-dimensional vision involve, in such cases as those of scaling for perspective and optic flow, their reduction to additive operations by the implicit logarithmic transformation of image coordinates. Expressions for such computations are derived and applied to illustrative examples of sensor design. The advantages of polar exponential arrays over X-Y rasters for binocular vision are noted to encompass the inference of range and three-dimensional position from local image velocity without knowledge of pixel location, provided that the relative velocity of the target and sensor are known by some other means.
A method for improving word recognition for people with age-related maculopathies, which cause a loss of central vision, is discussed. It is found that the use of individualized compensation filters based on an person's normalized contrast sensitivity function can improve word recognition for people with age-related maculopathies. It is shown that 27-70 pct more magnification is needed for unfiltered words compared to filtered words. The improvement in word recognition is positively correlated with the severity of vision loss.
Several techniques to perform static and dynamic load balancing techniques for vision systems are presented. These techniques are novel in the sense that they capture the computational requirements of a task by examining the data when it is produced. Furthermore, they can be applied to many vision systems because many algorithms in different systems are either the same, or have similar computational characteristics. These techniques are evaluated by applying them on a parallel implementation of the algorithms in a motion estimation system on a hypercube multiprocessor system. The motion estimation system consists of the following steps: (1) extraction of features; (2) stereo match of images in one time instant; (3) time match of images from different time instants; (4) stereo match to compute final unambiguous points; and (5) computation of motion parameters. It is shown that the performance gains when these data decomposition and load balancing techniques are used are significant and the overhead of using these techniques is minimal.
A local change in intensity (edge) is a characteristic that is preserved when an image is filtered through a bandpass filter. Primal sketch representations of images, using the bandpass-filtered data, have become a common process since Marr proposed his model for early human vision. Here, researchers move beyond the primal sketch extraction to the recovery of intensity and reflectance representations using only the bandpass-filtered data. Assessing the response of an ideal step edge to the Laplacian of Gaussian (NAb/A squared G) filter, they found that the resulting filtered data preserves the original change of intensity that created the edge in addition to the edge location. Using the filtered data, they can construct the primal sketches and recover the original (relative) intensity levels between the boundaries. It was found that the result of filtering an ideal step edge with the Intensity-Dependent Spatial Summation (IDS) filter preserves the actual intensity on both sides of the edge, in addition to the edge location. The IDS filter also preserves the reflectance ratio at the edge location. Therefore, one can recover the intensity levels between the edge boundaries as well as the (relative) reflectance representation. The recovery of the reflectance representation is of special interest as it erases shadowing degradations and other dependencies on temporal illumination. This method offers a new approach to low-level vision processing as well as to high data-compression coding. High compression can be gained by transmitting only the information associated with the edge location (edge primitives) that is necessary for the recovery
A top-down image processing scheme is described. A three-dimensional model of a robotic working environment, with robot manipulators, workpieces, cameras, and on-the-scene visual enhancements is employed to control and direct the image processing, so that rapid, robust algorithms act in an efficient manner to continually update the model. Only the model parameters are communicated, so that savings in bandwidth are achieved. This image compression by modeling is especially important for control of space telerobotics. The background for this scheme lies in an hypothesis of human vision put forward by the senior author and colleagues almost 20 years ago - the Scanpath Theory. Evidence was obtained that repetitive sequences of saccadic eye movements, the scanpath, acted as the checking phase of visual pattern recognition. Further evidence was obtained that the scanpaths were apparently generated by a cognitive model and not directly by the visual image. This top-down theory of human vision was generalized in some sense to the frame in artificial intelligence. Another source of the concept arose from bioengineering instrumentation for measuring the pupil and eye movements with infrared video cameras and special-purpose hardware.
All vision systems, both human and machine, transform the spatial image into a coded representation. Particular codes may be optimized for efficiency or to extract useful image features. Researchers explored image codes based on primary visual cortex in man and other primates. Understanding these codes will advance the art in image coding, autonomous vision, and computational human factors. In cortex, imagery is coded by features that vary in size, orientation, and position. Researchers have devised a mathematical model of this transformation, called the Hexagonal oriented Orthogonal quadrature Pyramid (HOP). In a pyramid code, features are segregated by size into layers, with fewer features in the layers devoted to large features. Pyramid schemes provide scale invariance, and are useful for coarse-to-fine searching and for progressive transmission of images. The HOP Pyramid is novel in three respects: (1) it uses a hexagonal pixel lattice, (2) it uses oriented features, and (3) it accurately models most of the prominent aspects of primary visual cortex. The transform uses seven basic features (kernels), which may be regarded as three oriented edges, three oriented bars, and one non-oriented blob. Application of these kernels to non-overlapping seven-pixel neighborhoods yields six oriented, high-pass pyramid layers, and one low-pass (blob) layer.
An ASEA IRB90 robotic manipulator with attached inspection cameras was moved through a Space Shuttle Payload Assist Module (PAM) Cradle under computer control. The Operator and Operator Control Station, including graphics simulation, gross-motion spatial planning, and machine vision processing, were located at JPL. The Safety and Support personnel, PAM Cradle, IRB90, and image acquisition system, were stationed at the Kennedy Space Center (KSC). Images captured at KSC were used both for processing by a machine vision system at JPL, and for inspection by the JPL Operator. The system found collision-free paths through the PAM Cradle, demonstrated accurate knowledge of the location of both objects of interest and obstacles, and operated with a communication delay of two seconds. Safe operation of the IRB90 near Shuttle flight hardware was obtained both through the use of a gross-motion spatial planner developed at JPL using artificial intelligence techniques, and infrared beams and pressure sensitive strips mounted to the critical surfaces of the flight hardward at KSC. The Demonstration showed that telerobotics is effective for real tasks, safe for personnel and hardware, and highly productive and reliable for Shuttle payload operations and Space Station external operations.
The principle function of vision is to measure the environment. As demonstrated by the coordination of motor actions with the positions and trajectories of moving objects in cluttered environments and by rapid recognition of solid objects in varying contexts from changing perspectives, vision provides real-time information about the geometrical structure and location of environmental objects and events. The geometric information provided by 2-D spatial displays is examined. It is proposed that the geometry of this information is best understood not within the traditional framework of perspective trigonometry, but in terms of the structure of qualitative relations defined by congruences among intrinsic geometric relations in images of surfaces. The basic concepts of this geometrical theory are outlined.