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.
LANDSAT signature development program (LSDP) automatically produces unsupervised classification of scene from LANDSAT data tape. Program is effective enough to be useful to sophisticated remote sensing analyst yet is simple enough to be utilized by ground truth investigators who have only basic understanding of computer and remote sensing procedures.
The author has identified the following results. The maximum likelihood classifier on the ERIPS failed to show any improvement in accuracy when comparing high-gain LANDSAT data with the simulated normal-gain data. Even if an improvement in accuracy had been detected, the timespan within the crop growing season when the use of high-gain data could be advantageous is limited. A total of six sets of LANDSAT-1 imagery is available for the analysis of gain effects. Three of the sets are high-gain data in MSS bands 4 and 5, whereas the other three are simulated normal-gain. The four sets obtained over Imperial Valley have good supporting ground truth information for training; the two sets over Kansas do not. The use of various combinations of the six data sets is recommended in order to uncover the possible advantages of using the data.
The SAGE mission is discussed along with the role of the Nimbus 7 experiment. Other topics discussed include: ground truth measurements, data collection and processing, SAGE instrumentation, and launch sequence.
The author has identified the following significant results. An accuracy of 90/85 was achieved with the October estimates which had a relative bias of -9.9 percent and a coefficient of variation of 5.2 percent for the total wheat production in the USGP. The probability was 0.9 that the LACIE estimate was within + or - 15 percent of true wheat production for the USGP. The LACIE spring wheat production underestimates in August, September, and October were the results of area underestimates for spring wheat in the USNGP region. The winter wheat blind study showed that the average proportion estimates were significantly different from the average dot-count, ground truth proportions at the USSGP and USGP-7 levels.
It is shown that low-altitude aerial photography, using specific interpretive techniques, can effectively delineate sea-grass beds, oyster beds, and other underwater features. Various techniques were used on several sets of aerial imagery. Imagery was tested using several data analysis methods, ground truth, and biological testing. Approximately 45,000 acres of grass beds, 2,500 acres of oyster beds, and 4,200 acres of dredged canals were mapped. This data represents selected sites only. Areas chosen have the highest quality water in Brevard County and are among the most highly recognized biologically productive waters in Florida.
The purpose of the experiment was to determine if remote microwave sensing of snowpack data could be used to predict runoff, thereby allowing more efficient management of the water supply. A four-frequency microwave radiometer system was attached to a truck-mounted aerial lift and was used to gather data on snowpacks at three different sites in the Colorado Rocky Mountains. Ground truth data measurements (density, temperature, grain size, hardness, and free-liquid water content) were taken at each site corresponding to each microwave scan.
This paper evaluates the costs of producing four remote sensing-based information products: timber volume estimate tables, Level II land use/land cover maps, soil maps, and vegetative cover maps. Two production methods for each product are evaluated, one is based on digital processing of satellite data, and the other on conventional photointerpretation of aircraft data. For each product, a comparison is conducted of the two strategies' production costs (including data acquisition, 'ground truthing', interpretation, compilation, and printing charges) and their performance (as measured by accuracy and timeliness). Each of the production methods reviewed has been demonstrated - either operationally or experimentally - and the costs, timeliness and other performance estimates presented are based on observations made in practice. The results show that for these products, satellite-based production results in significant cost and timeliness improvements at the cost of a loss in accuracy.
A study is made of the effects of quantization of the radar returns transmitted from aircraft or spacecraft employing a synthetic aperture radar system. The study is based on the output images obtained after one-bit, two-bit, and eight-bit quantizations and comparing the results to ground truth. In this way the degradation resulting from data or bandwidth reduction is determined. Quantization is evaluated in terms of crater scene, number of looks, and transmission error rate. It is found that two-bit quantization of raw radar data from homogeneous scenes processed to 32 looks yields nearly all the details of the original. One-bit quantization of raw radar data from homogeneous scenes processed to 32 looks yields a good visual representation of the scene but some fine detail is lost and the absolute reflectivity level is not reliable. Image quality is observed to improve with more looks and video and intermediate frequency quantization are not distinguishable even for one-bit quantizations. Image quality is not influenced by bit error rates less than about 2 to the -7th power.
The reported investigation is concerned with the use of Landsat remote sensing to define input parameters for an array of hydrologic models which are used to synthesize streamflow and water quality parameters in the planning or management process. The ground truth sampling and problems involved in translating the remotely sensed information into hydrologic model parameters are discussed. Questions related to the modification of existing models for compatibility with remote sensing capabilities are also examined. It is shown that the input parameters of many models are presently overdefined in terms of the sensitivity and accuracy of the model. When this overdefinition is recognized many of the models currently considered to be incompatible with remote sensing capabilities can be modified to make possible use with sensors having rather low resolutions.
The author has identified the following significant results. The green number and brightness scatter plots, channel plots of radiance values, and visual study of the imagery indicate separability between barley and spring wheat/oats during the wheat mid-heading to mid-ripe stages. In the LACIE Phase 3 North Dakota data set, the separation time is more specifically the wheat soft dough stage. At this time, the barley is ripening, and is therefore, less green and brighter than the wheat. Only 4 of the 18 segments studied indicate separation of barley/other spring small grain, even though 11 of the segments have acquisitions covering the wheat soft dough stage. The remaining seven segments had less than 5 percent barley based on ground truth data.
The spatial processing needs of LACIE include the location of agricultural test sites, and the registration of ground truth to LANDSAT imagery. The technological aspects of LACIE cartographic support, the need for cartography in satellite crop surveys, and proposed improvements which would enhance support of future programs are discussed.
The major research conducted during the three years of LACIE to solve problems associated with segment wheat area estimation is reviewed. Topics covered include proportion estimation, clustering, feature extraction, and signature extension. It would appear that LANDSAT-1 and LANDSAT-2 data do not contain enough information to discriminate between crop types perfectly all the time and, therefore, a basic problem arises when no ground truth data on crop types in the area are available. New approaches are needed to reduce labeling error. Perhaps better use of multiyear LANDSAT data, a more detailed understanding of the cropping practices in the area, better crop calendar prediction, and a better understanding of the limiting sources of error in LANDSAT data related to crop discrimination may provide the insight required to develop improved designs.
The author has identified the following signficant results. Examination of both machine estimates and stratified areal estimates produced by clustering and classification reveal no significant differences between the proportion estimates and ground truth estimates. When testing the variances of the machine estimates, a significant reduction in the variances was found when the number of starting dots was increased from 30 to 45. The variances were again reduced, although not significantly, when the number of starting dots was increased from 45 to 60. From these results, 60 starting dots are recommended for a three-category classifier. When examining the variances of the estimates for the four estimation procedures (using 60 dots), no significant differences were found between procedures. Thus, only the machine clustering may be used to produce an estimate, and the stratified areal estimate computations and maximum likelihood classification can be deleted.
The author has identified the following significant results. Ground truth of soil moisture content, and ambient air and soil temperatures were acquired concurrently with measurements of soil moisture in bare fields and fields covered with grass, corn, and soybeans obtained with 1.4 GHz and 5 GHz radiometers mounted on a truck. The biomass of the vegetation was sampled about once a week. The measured brightness temperatures over the bare fields were compared with those of radiative transfer model calculations using as inputs the acquired soil moisture and temperatures data with appropriate values of dielectric constants for soil-water mixtures. A good agreement was found between the calculated and measured results over 10 deg to 70 deg incident angles. The presence of vegetation reduced the sensitivity of soil moisture sensing. At 1.4 GHz the sensitivity reduction ranged from about 20% for 10 cm tall grassland cover to over 50 to 60% for the dense soybean field. At 5 GHz corresponding reduction in sensitivity ranged from approximately 70% to approximately 90%.
The interaction mechanisms responsible for the microwave backscattering and emission behavior of snow were investigated, and models were developed relating the backscattering coefficient (sigma) and apparent temperature (T) to the physical parameters of the snowpack. The microwave responses to snow wetness, snow water equivalent, snow surface roughness, and to diurnal variations were investigated. Snow wetness was shown to have an increasing effect with increasing frequency and angle of incidence for both active and passive cases. Increasing snow wetness was observed to decrease the magnitude sigma and increase T. Snow water equivalent was also observed to exhibit a significant influence sigma and T. Snow surface configuration (roughness) was observed to be significant only for wet snow surface conditions. Diurnal variations were as large as 15 dB for sigma at 35 GHz and 120 K for T at 37 GHz. Simple models for sigma and T of a snowpack scene were developed in terms of the most significant ground-truth parameters. The coefficients for these models were then evaluated; the fits to the sigma and T measurements were generally good. Finally, areas of needed additional observations were outlined and experiments were specified to further the understanding of the microwave-snowpack interaction mechanisms.
A simple method of estimating moisture content (W) of a bare soil from the observed brightness temperature (T sub B) at 1.4 GHz is discussed. The method is based on a radiative transfer model calculation, which has been successfully used in the past to account for many observational results, with some modifications to take into account the effect of surface roughness. Besides the measured T sub B's, the three additional inputs required by the method are the effective soil thermodynamic temperature, the precise relation between W and the smooth field brightness temperature T sub B and a parameter specifying the surface roughness characteristics. The soil effective temperature can be readily measured and the procedures of estimating surface roughness parameter and obtaining the relation between W and smooth field brightness temperature are discussed in detail. Dual polarized radiometric measurements at an off-nadir incident angle are sufficient to estimate both surface roughness parameter and W, provided that the relation between W and smooth field brightness temperature at the same angle is known. The method of W estimate is demonstrated with two sets of experimental data, one from a controlled field experiment by a mobile tower and the other, from aircraft overflight. The results from both data sets are encouraging when the estimated W's are compared with the acquired ground truth of W's in the top 2 cm layer. An offset between the estimated and the measured W's exists in the results of the analyses, but that can be accounted for by the presently poor knowledge of the relationship between W and smooth field brightness temperature for various types of soils. An approach to quantify this relationship for different soils and thus improve the method of W estimate is suggested.
Detailed gravity anomalies are computed from a combination of Geos 3 satellite altimeter and terrestrial gravity data using least-squares principles. The mathematical model used is based on the Stokes' equation modified for a nonglobal solution. Using Geos 3 data in the calibration area, the effects of several anomaly parameter configurations and data densities/distributions on the anomalies and their accuracy estimates are studied. The accuracy estimates for 1 deg x 1 deg mean anomalies from low density altimetry data are of the order of 4 mgal. Comparison of these anomalies with the terrestrial data and also with Rapp's data derived using collocation techniques show rms differences of 7.2 and 4.9 mgal, respectively. Indications are that the anomaly accuracies can be improved to about 2 mgal with high density data. Estimation of 30 in. x 30 in. mean anomalies indicates accuracies of the order of 5 mgal. Proper verification of these results will be possible only when accurate ground truth data become available.
The Accuracy Assessment System (AAS) of the Large Area Crop Inventory Experiment (LACIE) was responsible for determining the accuracy and reliability of LACIE estimates of wheat production, area, and yield, made at regular intervals throughout the crop season, and for investigating the various LACIE error sources, quantifying these errors, and relating them to their causes. Some results of using the AAS during the three years of LACIE are reviewed. As the program culminated, AAS was able not only to meet the goal of obtaining accurate statistical estimates of sampling and classification accuracy, but also the goal of evaluating component labeling errors. Furthermore, the ground-truth data processing matured from collecting data for one crop (small grains) to collecting, quality-checking, and archiving data for all crops in a LACIE small segment.