Rapid Constrained Object Motion Estimation based on Centroid Localization of Semantically Labeled Objects
Explore the source record for details and available documents.
SEARCH · Search NASA
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.
Explore the source record for details and available documents.
Explore the source record for details and available documents.
Algorithm processes synthetic-aperture-radar (SAR) returns into optimal estimates of ranges, pointing angles, and Doppler centroids. Derives radar-beam-pointing angles partly from SAR data. Exploits information available in overlapping regions between adjacent image bursts or, equivalently, between adjacent radar beams corresponding to adjacent image bursts. Range or Doppler centroid obtained from estimate of range or Doppler value at which intensity in one image burst equals that in adjacent burst.
A computer program (ALGEVAL) has been developed to simulate the position estimating behavior of a centroid estimator algorithm using data typical of optical point spread function data recorded by an area array detector. Typical results are shown of varying detector properties and optical point spread function types. The detector parameters currently available for study include read noise mean value, dark current mean value and spatial variation, charge transfer efficiency and point spread function location, saturation level, signal level and pixel size. The program is capable of calculating any order centroid using an array size from 2 x 2 to 15 x 15 pixels. The output of the program is either a performance map, histogram data or tabluar data. A number of further developments are recommended.
Paper discusses performances of two algorithms for resolution of ambiguity in estimated Doppler centroid frequency of echoes in synthetic-aperture radar. One based on range-cross-correlation technique, other based on multiple-pulse-repetition-frequency technique.
Known for their simplistic design and continuous detonation, rotating detonation engines (RDEs) constitute a majority of current pressure gain combustion (PGC) research efforts. Experimental RDE operation times have been continuously extended through the use of rig cooling techniques. As the window of observable behavior is expanded, and as the technology matures toward eventual integration within gas turbines, monitoring techniques must evolve to better match industrial diagnostics. High-speed image analysis techniques prove useful to capture and evaluate the unsteady detonation behavior within the RDE. Traditional image analysis techniques, however, require extensive processing times which prohibit simultaneous monitoring. To better address this problem, a computer vision object detection methodology is proposed to quickly detect individual detonation waves within a single down-axis image. Detonation waves are detected in individual images by the implemented computer vision method You Only Look Once (YOLO) object detection network. In order to detect detonation waves, the network must first be trained using RDE images of interest, for which each required phase of network development is outlined. Detection of waves is improved through proper treatment of the collected image set, variation of Intersection over Union (IoU) and confidence thresholding, and through a parametric study of annotation dimensions. Each detected wave is described by its location and rotational direction, and locations are tracked to calculate wave velocity across each frame, leading to a timestep resolution of 20 µs. Wave velocities are also calculated through a series of frames, leading to a suitable average velocity estimation using as few as 10 frames. Uncertainty analysis accounting for variation in camera framerate, pixel width and annotation centroid locations estimates a total uncertainty of ±4.3% for velocity calculations, using the smallest annotation boxes. This method offers great reductions in processing times, as a step toward real-time monitoring of detonation waves within an RDE. Improving on previous studies, this technique is impartial to wave modes not included in the original training set and calculates wave velocities independent of high-speed pressure data. The ability to isolate waves within predicted bounding boxes will likely facilitate analysis of pixel intensity variation as an estimation of wave strength in future work.
Known for their simplistic design and continuous detonation, rotating detonation engines (RDEs) constitute a majority of current pressure gain combustion (PGC) research efforts. Experimental RDE operation times have been continuously extended through the use of rig cooling techniques. As the window of observable behavior is expanded, and as the technology matures toward eventual integration within gas turbines, monitoring techniques must evolve to better match industrial diagnostics. High-speed image analysis techniques prove useful to capture and evaluate the unsteady detonation behavior within the RDE. Traditional image analysis techniques, however, require extensive processing times which prohibit simultaneous monitoring. To better address this problem, a computer vision object detection methodology is proposed to quickly detect individual detonation waves within a single down-axis image. Detonation waves are detected in individual images by the implemented computer vision method You Only Look Once (YOLO) object detection network. In order to detect detonation waves, the network must first be trained using RDE images of interest, for which each required phase of network development is outlined. Detection of waves is improved through proper treatment of the collected image set, variation of Intersection over Union (IoU) and confidence thresholding, and through a parametric study of annotation dimensions. Each detected wave is described by its location and rotational direction, and locations are tracked to calculate wave velocity across each frame, leading to a timestep resolution of 20 µs. Wave velocities are also calculated through a series of frames, leading to a suitable average velocity estimation using as few as 10 frames. Uncertainty analysis accounting for variation in camera framerate, pixel width and annotation centroid locations estimates a total uncertainty of ±4.3% for velocity calculations, using the smallest annotation boxes. This method offers great reductions in processing times, as a step toward real-time monitoring of detonation waves within an RDE. Improving on previous studies, this technique is impartial to wave modes not included in the original training set and calculates wave velocities independent of high-speed pressure data. The ability to isolate waves within predicted bounding boxes will likely facilitate analysis of pixel intensity variation as an estimation of wave strength in future work.
The Batuski-Burns list of candidate superclusters was searched for X-ray emission in the HEAO 1 A-2 data base. No significant flux in excess of integrated member cluster emission was detected in directions associated with estimated supercluster centroids; a 3 sigma upper limit to the mean individual excess flux is 5 x 10 to the -12th ergs/sq cm/s. Implications of this bound are discussed in terms of fainter member clusters, hot diffuse gas, and active and normal galaxies. Prospects of supercluster observations related to future X-ray missions are also outlined.
A new method for improving centroid accuracy, thereby pointing accuracy, is proposed.
We present a weak-lensing study of the cluster A520 based on Advanced Camera for Surveys (ACS) data. The excellent data quality provides a mean source density of ∼109 arcmin{sup –2}, which improves both resolution and significance of the mass reconstruction compared to a previous study based on Wide Field Planetary Camera 2 (WFPC2) images. We take care in removing instrumental effects such as the charge trailing due to radiation damage of the detector and the position-dependent point-spread function. This new ACS analysis confirms the previous claims that a substantial amount of dark mass is present between two luminous subclusters where we observe very little light. The centroid of the dark peak in the current ACS analysis is offset to the southwest by ∼1' with respect to the centroid from the WFPC2 analysis. Interestingly, this new centroid is in better agreement with the location where the X-ray emission is strongest, and the mass-to-light ratio estimated with this centroid is much higher (813 ± 78 M {sub ☉}/L {sub R☉}) than the previous value; the aperture mass with the WFPC2 centroid provides a consistent mass. Although we cannot provide a definite explanation for the dark peak, we discuss a revised scenario, wherein dark matter with a more conventional range (σ{sub DM}/m {sub DM} < 1 cm{sup 2} g{sup –1}) of self-interacting cross-section can lead to the detection of this dark substructure. If supported by detailed numerical simulations, this hypothesis opens up the possibility that the A520 system can be used to establish a lower limit of the self-interacting cross-section of dark matter.
We use the coadded spectra of 32 epochs of Sloan Digital Sky Survey (SDSS) Reverberation Mapping Project observations of 482 quasars with z > 1.46 to highlight systematic biases in the SDSS- and Baryon Oscillation Spectroscopic Survey (BOSS)-pipeline redshifts due to the natural diversity of quasar properties. We investigate the characteristics of this bias by comparing the BOSS-pipeline redshifts to an estimate from the centroid of He ii λ 1640. He ii has a low equivalent width but is often well-defined in high-S/N spectra, does not suffer from self-absorption, and has a narrow component which, when present (the case for about half of our sources), produces a redshift estimate that, on average, is consistent with that determined from [O ii] to within the He ii and [O ii] centroid measurement uncertainties. The large redshift differences of ∼1000 km s{sup −1}, on average, between the BOSS-pipeline and He ii-centroid redshifts, suggest there are significant biases in a portion of BOSS quasar redshift measurements. Adopting the He ii-based redshifts shows that C iv does not exhibit a ubiquitous blueshift for all quasars, given the precision probed by our measurements. Instead, we find a distribution of C iv-centroid blueshifts across our sample, with a dynamic range that (i) is wider than that previously reported for this line, and (ii) spans C iv centroids from those consistent with the systemic redshift to those with significant blueshifts of thousands of kilometers per second. These results have significant implications for measurement and use of high-redshift quasar properties and redshifts, and studies based thereon.
The Geoscience Lidar Altimeter System (GLAS) is a laser altimeter designed to measure temporal changes in the topography and mass balance of the Earth's ice sheets. GLAS is scheduled to be launched as part of the Earth Observing System (EOS) in 2001. Due to the large pulse width of GLAS and the slope (and roughness) of the ice sheets, the mean elevation of the laser's surface spot will be estimated from the centroid of the return pulse. While the accuracy of individual altimetry measurements is expected to be 15 cm, a cross-over technique that averages the elevation differences at selected points on the ice sheets will allow GLAS to detect mean ice elevation changes as small as 2 cm per year. One factor that may degrade the proposed accuracy of the altimeter measurements is the effect of multiple scattering by thin clouds and aerosols over the ice sheets. In the return pulse, some photons from the lidar are slightly deflected by cloud particles but still return to the instrument detector. Since these photons travel a longer path than photons that pass directly to and from the surface, the mean travel time of the pulse is lengthened and the centroid of the return pulse is shifted toward a later time . This study intends to determine the importance of this source of altimetry error in the GLAS measurements. Two methods will be used to estimate path delays in the GLAS lidar for a range of cloud conditions. An analytic expression of a doubly scattered return signal (Eloranta, 1972) will be compared to Monte Carlo simulations that estimate path delays by cloud, aerosol and molecular scattering. The results from both methods will be discussed in relationship to current knowledge of Arctic cloud climatology, and the impact of multiple scattering on GLAS altimeter measurements will be evaluated.
Hartmann testing of x-ray telescopes is a simple test method to retrieve and analyze alignment errors and low-order circumferential errors of x-ray telescopes and their components. A narrow slit is scanned along the circumference of the telescope in front of the mirror and the centroids of the images are calculated. From the centroid data, alignment errors, radius variation errors, and cone-angle variation errors can be calculated. Mean cone angle, mean radial height (average radius), and the focal length of the telescope can also be estimated if the centroid data is measured at multiple focal plane locations. In this paper we present the basic equations that are used in the analysis process. These equations can be applied to full circumference or segmented x-ray telescopes. We use the Optical Surface Analysis Code (OSAC) to model a segmented x-ray telescope and show that the derived equations and accompanying analysis retrieves the alignment errors and low order circumferential errors accurately.
Abstract We present an uncertainty analysis on correlation‐based time delay estimate, the basis for broadband lightning interferometry. A normal interferometry would yield much higher uncertainties than the theoretically predicted lower bound due to finite correlation window lengths. However, if the signals are aligned in time, the uncertainties approach the lower bound and can be used to indicate the interferometric uncertainties. Based on this, we introduce a beam steering interferometry technique. It first estimates a direction centroid for the lightning sources with a normal interferometry and computes the time delays among the signals. It then shifts the raw data with the time delays to align the signals roughly in time and reprocess the aligned signals. It finally pushes the reprocessed results to their correct positions, based on the time delays estimated in the first step. We apply this technique on a fast positive breakdown (FPB) process that started a normal intracloud lightning. The FPB process is shown to have a much more complex structure and development than a normal interferometry would provide. More importantly, from both interferometric and polarization analyses, the FPB appears to be ignited by a cosmic ray shower (CRS). We estimate the radio frequency strength and frequency content related to a presumed CRS in a thunderstorm electric field and find that they are in agreement with the observations. We examine the electric field effect of the CRS front and find that it could raise the field above the threshold for positive breakdown and is capable of igniting the FPB (and the lightning).
The centroid depths and source mechanisms of 12 large earthquakes on transform faults of the northern Mid-Atlantic Ridge were determined from an inversion of long-period body waveforms. The earthquakes occurred on the Gibbs, Oceanographer, Hayes, Kane, 15 deg 20 min, and Vema transforms. The depth extent of faulting during each earthquake was estimated from the centroid depth and the fault width. The source mechanisms for all events in this study display the strike slip motion expected for transform fault earthquakes; slip vector azimuths agree to 2 to 3 deg of the local strike of the zone of active faulting. The only anomalies in mechanism were for two earthquakes near the western end of the Vema transform which occurred on significantly nonvertical fault planes. Secondary faulting, occurring either precursory to or near the end of the main episode of strike-slip rupture, was observed for 5 of the 12 earthquakes. For three events the secondary faulting was characterized by reverse motion on fault planes striking oblique to the trend of the transform. In all three cases, the site of secondary reverse faulting is near a compression jog in the current trace of the active transform fault zone. No evidence was found to support the conclusions of Engeln, Wiens, and Stein that oceanic transform faults in general are either hotter than expected from current thermal models or weaker than normal oceanic lithosphere.
The centroid depths and source mechanisms of 12 large earthquakes on transform faults of the northern Mid-Atlantic Ridge were determined from an inversion of long-period body waveforms. The earthquakes occurred on the Gibbs, Oceanographer, Hayes, Kane, 15 deg 20 min, and Vema transforms. The depth extent of faulting during each earthquake was estimated from the centroid depth and the fault width. The source mechanisms for all events in this study display the strike slip motion expected for transform fault earthquakes; slip vector azimuths agree to 2 to 3 deg of the local strike of the zone of active faulting. The only anomalies in mechanism were for two earthquakes near the western end of the Vema transform which occurred on significantly nonvertical fault planes. Secondary faulting, occurring either precursory to or near the end of the main episode of strike-slip rupture, was observed for 5 of the 12 earthquakes. For three events the secondary faulting was characterized by reverse motion on fault planes striking oblique to the trend of the transform. In all three cases, the site of secondary reverse faulting is near a compression jog in the current trace of the active transform fault zone. No evidence was found to support the conclusions of Engeln, Wiens, and Stein that oceanic transform faults in general are either hotter than expected from current thermal models or weaker than normal oceanic lithosphere.
Purpose: As it is generally difficult to outline the pancreas on an ultrasound b-mode image, visualized structures such as the portal or the splenic veins are assumed to have the same motion as the pancreas. These structures can be used as a surrogate for monitoring pancreas motion during radiation therapy (RT) delivery using ultrasound. To verify this assumption, we studied the motion difference between the head of the pancreas, the portal vein, the tail of the pancreas, and splenic vein. Methods: 4DCT data acquired during RT simulation were analyzed for a total of 5 randomly selected patients with pancreatic cancer. The data was sorted into 10 respiratory phases from 0% to 90% (0%: end of the inspiration, 50%: end of expiration) . The head of the pancreas (HP), tail of the pancreas (TP), portal vein (PV), and splenic vein (SV) were contoured on all 10 phases. The volume change and motion were measured in the left-right (LR), anterior-superior (AP), and superior-inferior (SI) directions. Results: The volume change for all patients/phases were: 1.2 ± 3% for HP, 0.78 ± 1.6% for PV, 2.5 ± 2.9% for TP, and 0.53 ± 2.1% for SV. Motion for each structure was estimated from the centroid displacements due to the uniformity of the structures and the small volume change. The measured motion between HP and PV was: LR: 0.1 ± 0.17 mm, AP: 0.04 ± 0.1 mm, SI: 0.17 ± 0.16 mm and between TP and the PV was: LR: 0.05 ± 0.3 mm, AP: 0.1 ± 0.4 mm, SI: 0.01 ± 0.022 mm. Conclusion: There are small motion differences between the portal vein and the head of the pancreas, and the splenic vein and the tail of the pancreas. This suggests the feasibility of utilizing these features for monitoring the pancreas motion during radiation therapy.
The feasibility of using long wavelength Earth thermal (infrared) images for telescope tracking/pointing application. for both Deep Space Free- pace Optical Communications has been investigated and is reported her. The advantage of this technology rests on using full Earth images in this band, which yield more accurate estimates of geometric centroids than that of Earth images in the visible band. Another major advantage is that these images are nearly independent of Earth phase angle. The results of the study show that at a Mars range, with currently available sensors, a noise equivalent angle of 10 to 150 nanoradians and a bias error of better than 80 nanoradians can be obtained. This enables precise pointing of the optical communications beam for high data rate links.