ecospec v0.1.0
Bespoke software for segmenting plants in ecoFABs to monitor growth and health. Contains python libraries for image alignment, training neural networks and running inference.
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.
Bespoke software for segmenting plants in ecoFABs to monitor growth and health. Contains python libraries for image alignment, training neural networks and running inference.
Hyperspectral imaging provides a powerful tool for analyzing above-ground plant characteristics in fabricated ecosystems, offering rich spectral information across diverse wavelengths. This study presents an efficient workflow for hyperspectral data segmentation and subsequent data analytics, minimizing the need for user annotation through the use of ensembles of sparse mixed scale convolution neural networks. The segmentation process leverages the diversity of ensembles to achieve high accuracy with minimal labeled data, reducing labor-intensive annotation efforts. To further enhance robustness, we incorporate image alignment techniques to address spatial variability in the dataset. Downstream analysis focuses on using the segmented data for processing spectral data, enabling monitoring of plant health. This approach provides a scalable solution for spectral segmentation, and facilitates actionable insights into plant conditions in complex, controlled environments. Our results demonstrate the utility of combining advanced machine learning techniques with hyperspectral analytics for high-throughput plant monitoring.
The Broad Band X-Ray Telescope (BBXRT) experiment has been designed for high sensitivity, moderate resolution spectrophotometry of cosmic sources in the range 0.5 to 12 keV. Principal elements are two co-aligned imaging telescopes with cooled Si(Li) detectors at each focus. The mirror design uses an approximate geometry of tightly nested, conical, foil reflectors that allow a large filling factor (high throughput) at small grazing angles (high energy response). Each detector consists of a cluster of five elements defined with grooves on a single crystal. This arrangement affords some spatial resolution as well as a means of substantial background reduction. Shuttle borne measurements will typically be of 2000 s duration for sources with a flux of 10 to the -12 erg/sq cm-s.
An analysis was performed on the P-format computer compatible tapes for the Landsat-4 Thematic Mapper (TM) scene 40109-15140 acquired November 2, 1982 over Washington, D.C. and environs. Three tests of sensor geometry were undertaken: (1) band-to-band registration, (2) line-to-line registration between swaths, and (3) geometric correction of sensor and spacecraft/ephemeris characteristics. The band-to-band and line-to-line registration was measured at one hundred pixel spacings along a line using the phase correlation image alignment method developed by Kuglin and Hines and adapted to a one-dimensional FFT correlation technique. The sensor and spacecraft geometric calibration analysis was checked by identifying 75 ground control points in the TM scene and on 1:24,000 topographic maps. A least squares fit of the gaps was computed and the vector offsets of the residuals was tabulated and plotted.
A review of Doppler Global Velocimetry (DGV) data obtained during wind tunnel tests on a 75 degree swept delta wing was performed. High frequency variations observed in normalized data files are attributed to image alignment problems. Unfortunately, initial DGV velocity data compared poorly with baseline reference data. Nonlinear DGV system operation during the tests is the likely source of this problem. Corrected data compares much more favorably and suggests that DGV is a valid measurement technique. Future DGV investigations should include a method or means for monitoring laser frequency relative to the ALF transfer function behavior.
Optical testing is one of the most vital elements in the process of preparing an optical instrument for launch. Without well understood, well controlled, and well documented test procedures, current and future mission goals will be jeopardized. We should keep in mind that the reason we test is to provide an opportunity to catch errors, oversights, and problems on the ground, where solutions are possible and difficulties can be rectified. Consequently, it is necessary to create tractable test procedures that truly provide a measure of the performance of all optical elements and systems under conditions which are close to those expected in space. Where testing is not feasible, accurate experiments are required in order to perfect models that can exactly predict the optical performance. As we stretch the boundaries of technology to perform more complex space and planetary investigations, we must expand the technology required to test the optical components and systems which we send into space. As we expand the observational wavelength ranges, so must we expand our range of optical sources and detectors. As we increase resolution and sensitivity, our understanding of optical surfaces to accommodate more stringent figure and scatter requirements must expand. Only with research and development in these areas can we hope to achieve success in the ever increasing demands made on optical testing by the highly sophisticated missions anticipated over the next two decades. Technology assessment and development plan for surface figure, surface roughness, alignment, image quality, radiometric quantities, and stray light measurement are presented.
Previous research has demonstrated that a Head-Up Display (HUD) can be used to enable more capacity and safer aircraft surface operations. This previous research also noted that the HUD exhibited two major limitations which hindered the full potential of the display concept: 1) the monochrome HUD format; and, 2) a limited, fixed field of regard. Full-color Head Worn Displays (HWDs) with very small sizes and weights are emerging to the extent that this technology may be practical for commercial and business aircraft operations. By coupling the HWD with a head tracker, full-color, out-the-window display concepts with an unlimited field-of-regard may be realized to improve efficiency and safety in surface operations. A ground simulation experiment was conducted at NASA Langley to evaluate the efficacy of head-worn display applications which may directly address the limitations of the HUD while retaining all of its advantages in surface operations. The simulation experiment used airline crews to evaluate various displays (HUD, HWD) and display concepts in an operationally realistic environment by using a Chicago, O Hare airport database. The results pertaining to the implications of HWDs for commercial business and transport aviation applications are presented herein. Overall HWD system latency was measured and found to be acceptable, but not necessarily optimal. A few occurrences of simulator sickness were noted while wearing the HWD, but overall there appears to be commercial pilot acceptability and usability to the concept. Many issues were identified which need to be addressed in future research including continued reduction in user encumbrance due to the HWD, and improvement in image alignment, accuracy, and boresighting.
We present a second epoch of Chandra observations of the Type Ia Large Magellanic Cloud supernova remnant(SNR) 0509-68.7 (N103B) obtained in 2017. When combined with the earlier observations from 1999, we have a17.4 year baseline with which we can search for evidence of the remnant's expansion. Although the lack of strongpoint source detections makes absolute image alignment at the necessary accuracy impossible, we can measure thechange in the diameter and the area of the remnant, and find that it has expanded by an average velocity of 4170(2860, 5450) km s1. This supports the picture of this being a young remnant; this expansion velocity correspondsto an undecelerated age of 850 years, making the real age somewhat younger, consistent with results from lightecho studies. Previous infrared observations have revealed high densities in the western half of the remnant, likelyfrom circumstellar material, so it is probable that the real expansion velocity is lower on that side of the remnantand higher on the eastern side. A similar scenario is seen in Kepler's SNR. N103B joins the rare class ofMagellanic Cloud SNRs with measured proper motions.
Phase retrieval requires calculation of the real-valued phase of the pupil fimction from the image intensity distribution and characteristics of an optical system. Genetic 'algorithms were used to solve two one-dimensional phase retrieval problem. A GA successfully estimated the coefficients of a polynomial expansion of the phase when the number of coefficients was correctly specified. A GA also successfully estimated the multiple p h e s of a segmented optical system analogous to the seven-mirror Systematic Image-Based Optical Alignment (SIBOA) testbed located at NASA s Marshall Space Flight Center. The SIBOA testbed was developed to investigate phase retrieval techniques. Tiphilt and piston motions of the mirrors accomplish phase corrections. A constant phase over each mirror can be achieved by an independent tip/tilt correction: the phase Conection term can then be factored out of the Discrete Fourier Tranform (DFT), greatly reducing computations.
A reduced surrogate, one point crossover genetic algorithm with random rank-based selection was used successfully to estimate the multiple phases of a segmented optical system modeled on the seven-mirror Systematic Image-Based Optical Alignment testbed located at NASA's Marshall Space Flight Center.
We demonstrate that measurements of the gravitational tidal field made with spectroscopic redshifts can be improved with information from imaging surveys. The average orientation of small groups of galaxies, or “multiplets” is correlated with large-scale structure and is used to measure the direction of tidal forces. Previously, multiplet intrinsic alignment has been measured in DESI using galaxies that have spectroscopic redshifts. The DESI Legacy Imaging catalog can be used to supplement multiplet catalogs. Our findings show that galaxy positions from the imaging catalog produce a measurement similar to the measurements made with only spectroscopic data. This demonstrates that imaging can improve our signal-to-noise ratio for multiplet alignment in DESI.
Data from images produced by slightly incorrectly figured concave primary mirror in telescope processed into estimate of spherical aberration of mirror, by use of algorithm finding nonlinear least-squares best fit between actual images and synthetic images produced by multiparameter mathematical model of telescope optical system. Estimated spherical aberration, in turn, converted into estimate of deviation of reflector surface from nominal precise shape. Algorithm devised as part of effort to determine error in surface figure of primary mirror of Hubble space telescope, so corrective lens designed. Modified versions of algorithm also used to find optical errors in other components of telescope or of other optical systems, for purposes of testing, alignment, and/or correction.
A paper describes an optical stimulus that produces more consistent results, and can be automated for unattended, routine generation of data analysis products needed by the integration and testing team assembling a high-fidelity imaging spectrometer system. One key attribute of the system is an arrangement of pick-off mirrors that provides multiple input beams (five in this implementation) to simultaneously provide stimulus light to several field angles along the field of view of the sensor under test, allowing one data set to contain all the information that previously required five data sets to be separately collected. This stimulus can also be fed by quickly reconfigured sources that ultimately provide three data set types that would previously be collected separately using three different setups: Spectral Response Function (SRF), Cross-track Response Function (CRF), and Along-track Response Function (ARF), respectively. This method also lends itself to expansion of the number of field points if less interpolation across the field of view is desirable. An absolute minimum of three is required at the beginning stages of imaging spectrometer alignment.
The space astrometry mission Gaia will construct a dense optical QSO-based celestial reference frame. For consistency between optical and radio positions, it will be important to align the Gaia and VLBI frames (International Celestial Reference Frame) with the highest accuracy. In this respect, it is found that only 10% of the ICRF sources are suitable to establish this link (70 sources), either because most of the ICRF sources are not bright enough at optical wavelengths or because they show extended radio emission which precludes reaching the highest astrometric accuracy. In order to improve the situation, we initiated a multi-step VLBI observational project, dedicated to finding additional suitable radio sources for aligning the two frames. The sample consists of about 450 optically-bright radio sources, typically 20 times weaker than the ICRF sources, which have been selected by cross-correlating optical and radio catalogs. The initial observations, aimed at checking whether these sources are detectable with VLBI, and conducted with the European VLBI Network (EVN) in 2007, showed an excellent 90% detection rate. This paper reports on global VLBI observations carried out in March 2008 to image 105 from the 398 previously detected sources. All sources were successfully imaged, revealing compact VLBI structure for about half of them, which is very promising for the future.
This presentation will review the objectives and current status of two Small Business Innovative Research being performed by the Sirius Group, under the direction of MSFC. They all relate to the development of advanced optical systems technologies for automated segmented mirror alignment techniques and fundamental design methodologies for ultralight structures. These are important to future astronomical missions in space.
Explore the source record for details and available documents.
We present a design and tolerancing approach that permits the achievement of a high degree of spatial and spectral uniformity of response from a pushbroom imaging spectrometer. Such uniformity of response is crucial for the extraction of accurate spectroscopic information from remotely sensed data. The spectrometer system example comprises two independent spectrometer modules covering the 400-2500 nm range, seperated through a dichroic mirror.
In this paper, we demonstrate a simple algorithm that projects low resolution (LR) images differing in subpixel shifts on a high resolution (HR) also called super resolution (SR) grid. The algorithm is very effective in accuracy as well as time efficiency. A number of spatial interpolation techniques using nearest neighbor, inverse-distance weighted averages, Radial Basis Functions (RBF) etc. used in projection yield comparable results. For best accuracy of reconstructing SR image by a factor of two requires four LR images differing in four independent subpixel shifts. The algorithm has two steps: i) registration of low resolution images and (ii) shifting the low resolution images to align with reference image and projecting them on high resolution grid based on the shifts of each low resolution image using different interpolation techniques. Experiments are conducted by simulating low resolution images by subpixel shifts and subsampling of original high resolution image and the reconstructing the high resolution images from the simulated low resolution images. The results of accuracy of reconstruction are compared by using mean squared error measure between original high resolution image and reconstructed image. The algorithm was tested on remote sensing images and found to outperform previously proposed techniques such as Iterative Back Projection algorithm (IBP), Maximum Likelihood (ML), and Maximum a posterior (MAP) algorithms. The algorithm is robust and is not overly sensitive to the registration inaccuracies.