Multiscale Image Decomposition and Analysis of Shapes [Slides]
Abstract not provided.
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.
Abstract not provided.
Abstract not provided.
Explore the source record for details and available documents.
Bioaerosol and soil particles are ubiquitous in the environment. They are multicomponent and complex in nature displaying mixed inorganic and organic components. The way components are mixed in a bioaerosol sample is referred to as its mixing state. Soil particles are also a mixture of inorganic (mineral) and organic (soil organic matter) components. Bioaerosol particles contribute to a major fraction of coarse mode atmospheric particles, especially in the tropical areas, contributing up to 80 % of the particle mass concentration. The mixing state of particles is crucial to evaluate because it impacts several important environmental processes such as warm and cold cloud formation and radiation budget. Mixing states in aerosols are accompanied by chemical reactions across solid-liquid-gas interfaces. In this study, we utilized elemental compositions and microcopy images of thousands of atmospheric particles acquired by computer-controlled scanning electron microscope equipped with an energy-dispersive x-ray spectrometer to compute the mixing state of atmospheric particles. A 2D convolutional neural network (CNN), also known as convnet, was used to model the relationship between low resolution imaging data and higher resolution spectroscopy data, with the former as training input and the latter as target output. Two types of CNNs were implemented and tested; a basic CNN and an Inception-v3 network. For binary classification, the basic CNN achieved an accuracy of 84.29 % across all atom types, and the Inception-v3-like network achieved an accuracy 85.51 %. This study demonstrates the applicability of deep learning to handle large amounts of imaging/chemical spectroscopy data efficiently and evaluate particle mixing state from a range of environmental samples.
Abstract not provided
Explore the source record for details and available documents.
A mathematical model derived from the equation of heat conduction was developed to assist in interpreting thermal infrared images acquired from aircraft and spacecraft. The model assumes steady state boundary conditions. It contains parameters of rock and soil properties, atmospheric effects, site location, and season. The results predicted provide an explanation for the thermal differences among granite, limestone, and dolomite recorded in the December 1968 daytime and predawn flights over the Mill Creek, Oklahoma test site, during which representative thermal inertia and albedo values were used. A second test of the model made use of data acquired during the June 1970 predawn overflight of Mill Creek. A simple model of transient heating of the ground was constructed as an extension of the overall model, in order to examine the effects of atmospheric perturbations. The results obtained are consistent with those of ground observations made at the time of the overflight.
Simulation studies of selected aircraft and spacecraft acquired images were initiated in response to anticipated desires of the user community for simulated imagery prior to launch of ERTS-A. Principal modes of simulation included: (1) areas of coverage comparable to ERTS; (2) spatial resolutions within these images similar to ERTS; (3) spectral responses analogous to that expected from each channel on both the return beam vidicon and multispectral scanner, leading to production of photographic images that should appear similar to those representing each band in the two sensors; and (4) runthroughs of several analytical techniques, such as color density slicing, color additive viewing, and computer-generated reflectance and surface temperature maps, by which ERTS data can be analyzed, interpreted, and applied. The two areas involved in the simulation study were Wyoming and the Chesapeake Bay region.
Analysis during the quarter was carried out on geometric, radiometric, and information content aspects of both MSS and thematic mapper (TM) data. Test sites in Webster County, Iowa and Chicago, IL., and near Joliet, IL were studied. Band to band registration was evaluated and TM Bands 5 and 7 were found to be approximately 0.5 pixel out of registration with 1,2,3,4, and the thermal was found to be misregistered by 4 30 m pixels to the east and 1 pixel south. Certain MSS bands indicated nominally .25 pixel misregistration. Radiometrically, some striping was observed in TM bands and significant oscillatory noise patterns exist in MSS data which is possibly due to jitter. Information content was compared before and after cubic convolution resampling and no differences were observed in statistics or separability of basic scene classes.
Seven heterogeneous areas within the entire Des Moines, Iowa test site were selected to define candidate spectral training classes using a clustering algorithm. In addition to the 91 cluster nonsupervised classes, three supervised training classes were defined. The original candidate training classes were reduced to 42 spectrally separable training classes. The minimum and average transformed divergence values for the 42 spectral classes and for the best subsets of Y TM spectral bands are shown in a table. The best spectral band for any combination of 1 through 7 bands is the first middle IR band. The next best band is the near IR, followed by the red band and than the thermal IR. The best combination of four bands includes one from each of the four regions of the spectrum (visible, near IR, middle IR, and thermal IR).
Seven heterogeneous areas within the Des Moines, Iowa area test site were selected to define candidate spectral training classes using a clustering algorithm. In addition to the 91 cluster (nonsupervised) classes, three supervised training classes were defined and subsequently included in the training statistics file. The identity of all 94 candidate classes were determined using available reference data. Through analysis of the interclass separabilities, the original 94 candidate training classes were reduced to 42 spectrally separable final classes. The minimum average transformed divergence values for the 42 spectral classes and for the best subsets of TM spectral bands are shown in a table.
A comparative analysis of TM and MSS data was completed and the results indicate that there are half as many separable spectral classes in the MSS data than in TM. In addition, the minimum separability between classes was also much less in MSS data. Radiometric data quality was also investigated for the TM by computing power spectrum estimates for dark-level data from Lake Michigan. Two significant coherent noise frequencies were observed, one with a wavelength of 3.12 pixels and the other with a 17 pixel wavelength. The amplitude was small (nominally .6 digital count standard deviation) and the noise appears primarily in Bands 3 and 4. No significant levels were observed in other bands. Scan angle dependent brightness effects were also evaluated.
Classification performance from LANDSAT 4 TM and MSS data is evaluated using the SECHO computer program. The data accuracy is compared using forest, corn, soybeans, bare soil, grass, water, and urban areas as classes for investigation.
Reformatting software to handle LANDSAT 5 data in quadrant format was completed and tested. The sensor two-dimensional point spread function was estimated from scene data. Budget recalculations are discussed. Two publications done under this contract are named.
LANDSAT-4/5 data quality analysis was covered. Focus was on estimation of two-dimensional point-spread function estimation. A brief description is included.
A LANDSAT Thematic Mapper (TM) quality evaluation study was conducted to identify geometric and radiometric sensor errors in the post-launch environment. The study began with the launch of LANDSAT-4. Several error conditions were found, including band-to-band misregistration and detector-to detector radiometric calibration errors. Similar analysis was made for the LANDSAT-5 Thematic Mapper and compared with results for LANDSAT-4. Remaining band-to-band misregistration was found to be within tolerances and detector-to-detector calibration errors were not severe. More coherent noise signals were observed in TM-5 than in TM-4, although the amplitude was generally less. The scan direction differences observed in TM-4 were still evident in TM-5. The largest effect was in Band 4 where nearly a one digital count difference was observed. Resolution estimation was carried out using roads in TM-5 for the primary focal plane bands rather than field edges as in TM-4. Estimates using roads gave better resolution. Thermal IR band calibration studies were conducted and new nonlinear calibration procedures were defined for TM-5. The overall conclusion is that there are no first order errors in TM-5 and any remaining problems are second or third order.
LANDSAT-4 Thematic Mapper (TM) data performance and utility characteristics from an energy research and technology perspective is evaluated. The program focuses on evaluating applicational implications of using such data, in combination with other digital data, for current and future energy research and technology activities. Prime interest is in using TM data for siting, developing and operating federal energy facilities. Secondary interests involve the use of such data for resource exploration, environmental monitoring and basic scientific initiatives such as in support of the Continental Scientific Drilling Program.