Publications

Abstract not provided.

A method for tracking streaking targets (targets whose signatures are spread across multiple pixels in a focal plane array) is developed. The outputs of a bank of matched filters are thresholded and then used for measurement extraction. The use of the Deep Target Extractor (DTE, previously called the MLPMHT) allows for tracking in the very low observable (VLO) environment common when a streaking target is present. A definition of moving target signal to noise ratio (MT-SNR) is also presented as a metric for trackability. The extraction algorithm and the DTE are then tested across several variables, including trajectory, MT-SNR, and streak length. The DTE and measurement extraction process performs remarkably well in this difficult tracking environment on these data features.

Low signal-to-noise data processing algorithms for improved detection, tracking, discrimination and situational threat assessment are a key research challenge. As sensor technologies progress, the number of pixels will increase significantly. This will result in increased resolution, which could improve object discrimination, but unfortunately, will also result in a significant increase in the number of potential targets to track. Many tracking techniques, like multi-hypothesis trackers, suffer from a combinatorial explosion as the number of potential targets increase. As the resolution increases, the phenomenology applied towards detection algorithms also changes. For low resolution sensors, blob tracking is the norm. For higher resolution data, additional information may be employed in the detection and classification steps. The most challenging scenarios are those where the targets cannot be fully resolved, yet must be tracked and distinguished for neighboring closely spaced objects. Tracking vehicles in an urban environment is an example of such a challenging scenario. This report evaluates several potential tracking algorithms for large-scale tracking in an urban environment. The algorithms considered are: random sample consensus (RANSAC), Markov chain Monte Carlo data association (MCMCDA), tracklet inference from factor graphs, and a proximity tracker. Each algorithm was tested on a combination of real and simulated data and evaluated against a common set of metrics.

Abstract not provided.

Low signal-to-noise data processing algorithms for improved detection, tracking, discrimination and situational threat assessment are a key research challenge. As sensor technologies progress, the number of pixels will increase signi cantly. This will result in increased resolution, which could improve object discrimination, but unfortunately, will also result in a significant increase in the number of potential targets to track. Many tracking techniques, like multi-hypothesis trackers, suffer from a combinatorial explosion as the number of potential targets increase. As the resolution increases, the phenomenology applied towards detection algorithms also changes. For low resolution sensors, "blob" tracking is the norm. For higher resolution data, additional information may be employed in the detection and classfication steps. The most challenging scenarios are those where the targets cannot be fully resolved, yet must be tracked and distinguished for neighboring closely spaced objects. Tracking vehicles in an urban environment is an example of such a challenging scenario. This report evaluates several potential tracking algorithms for large-scale tracking in an urban environment.

Abstract not provided.

Abstract not provided.

A considerable amount research is being conducted on microalgae, since microalgae are becoming a promising source of renewable energy. Most of this research is centered on lipid production in microalgae because microalgae produce triacylglycerol which is ideal for biodiesel fuels. Although we are interested in research to increase lipid production in algae, we are also interested in research to sustain healthy algal cultures in large scale biomass production farms or facilities. The early detection of fluctuations in algal health, productivity, and invasive predators must be developed to ensure that algae are an efficient and cost-effective source of biofuel. Therefore we are developing technologies to monitor the health of algae using spectroscopic measurements in the field. To do this, we have proposed to spectroscopically monitor large algal cultivations using LIDAR (Light Detection And Ranging) remote sensing technology. Before we can deploy this type of technology, we must first characterize the spectral bio-signatures that are related to algal health. Recently, we have adapted our confocal hyperspectral imaging microscope at Sandia to have two-photon excitation capabilities using a chameleon tunable laser. We are using this microscope to understand the spectroscopic signatures necessary to characterize microalgae at the cellular level prior to using these signatures to classify the health of bulk samples, with the eventual goal of using of LIDAR to monitor large scale ponds and raceways. By imaging algal cultures using a tunable laser to excite at several different wavelengths we will be able to select the optimal excitation/emission wavelengths needed to characterize algal cultures. To analyze the hyperspectral images generated from this two-photon microscope, we are using Multivariate Curve Resolution (MCR) algorithms to extract the spectral signatures and their associated relative intensities from the data. For this presentation, I will show our two-photon hyperspectral imaging results on a variety of microalgae species and show how these results can be used to characterize algal ponds and raceways.

Line of sight jitter in staring sensor data combined with scene information can obscure critical information for change analysis or target detection. Consequently before the data analysis, the jitter effects must be significantly reduced. Conventional principal component analysis (PCA) has been used to obtain basis vectors for background estimation; however PCA requires image frames that contain the jitter variation that is to be modeled. Since jitter is usually chaotic and asymmetric, a data set containing all the variation without the changes to be detected is typically not available. An alternative approach, Scene Kinetics Mitigation, first obtains an image of the scene. Then it computes derivatives of that image in the horizontal and vertical directions. The basis set for estimation of the background and the jitter consists of the image and its derivative factors. This approach has several advantages including: (1) only a small number of images are required to develop the model, (2) the model can estimate backgrounds with jitter different from the input training images, (3) the method is particularly effective for sub-pixel jitter, and (4) the model can be developed from images before the change detection process. In addition the scores from projecting the factors on the background provide estimates of the jitter magnitude and direction for registration of the images. In this paper we will present a discussion of the theoretical basis for this technique, provide examples of its application, and discuss its limitations.

Abstract not provided.

With the continuing development of more capable data gathering sensors, comes an increased demand on the bandwidth for transmitting larger quantities of data. To help counteract that trend, a study was undertaken to determine appropriate lossy data compression strategies for minimizing their impact on target detection and characterization. The survey of current compression techniques led us to the conclusion that wavelet compression was well suited for this purpose. Wavelet analysis essentially applies a low-pass and high-pass filter to the data, converting the data into the related coefficients that maintain spatial information as well as frequency information. Wavelet compression is achieved by zeroing the coefficients that pertain to the noise in the signal, i.e. the high frequency, low amplitude portion. This approach is well suited for our goal because it reduces the noise in the signal with only minimal impact on the larger, lower frequency target signatures. The resulting coefficients can then be encoded using lossless techniques with higher compression levels because of the lower entropy and significant number of zeros. No significant signal degradation or difficulties in target characterization or detection were observed or measured when wavelet compression was applied to simulated and real data, even when over 80% of the coefficients were zeroed. While the exact level of compression will be data set dependent, for the data sets we studied, compression factors over 10 were found to be satisfactory where conventional lossless techniques achieved levels of less than 3.

Abstract not provided.

The combination of hyperspectral confocal fluorescence microscopy and multivariate curve resolution (MCR) provides an ideal system for improved quantitative imaging when multiple fluorophores are present. However, the presence of multiple noise sources limits the ability of MCR to accurately extract pure-component spectra when there is high spectral and/or spatial overlap between multiple fluorophores. Previously, MCR results were improved by weighting the spectral images for Poisson-distributed noise, but additional noise sources are often present. We have identified and quantified all the major noise sources in hyperspectral fluorescence images. Two primary noise sources were found: Poisson-distributed noise and detector-read noise. We present methods to quantify detector-read noise variance and to empirically determine the electron multiplying CCD (EMCCD) gain factor required to compute the Poisson noise variance. We have found that properly weighting spectral image data to account for both noise sources improved MCR accuracy. In this paper, we demonstrate three weighting schemes applied to a real hyperspectral corn leaf image and to simulated data based upon this same image. MCR applied to both real and simulated hyperspectral images weighted to compensate for the two major noise sources greatly improved the extracted pure emission spectra and their concentrations relative to MCR with either unweighted or Poisson-only weighted data. Thus, properly identifying and accounting for the major noise sources in hyperspectral images can serve to improve the MCR results. These methods are very general and can be applied to the multivariate analysis of spectral images whenever CCD or EMCCD detectors are used. Copyright © 2008 John Wiley & Sons, Ltd.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Abstract not provided.

A numerical model of the ESR process was used to study the effect of the various process parameters on the resulting temperature profiles, flow field, and pool shapes. The computational domain included the slag and ingot, while the electrode, crucible, and cooling water were considered as external boundary conditions. The model considered heat transfer, fluid flow, solidification, and electromagnetic effects. The predicted pool profiles were compared with experimental results obtained over a range of processing parameters from an industrial-scale 718 alloy ingot. The shape of the melt pool was marked by dropping nickel balls down the annulus of the crucible during melting. Thermocouples placed in the electrode monitored the electrode and slag temperature as melting progressed. The cooling water temperature and flow rate were also monitored. The resulting ingots were sectioned and etched to reveal the ingot macrostructure and the shape of the melt pool. Comparisons of the predicted and experimentally measured pool profiles show excellent agreement. The effect of processing parameters, including the slag cap thickness, on the temperature distribution and flow field are discussed. The results of a sensitivity study of thermophysical properties of the slag are also discussed.

Abstract not provided.

A particularly challenging problem associated with vacuum arc remelting occurs when trying to maintain accurate control of electrode melt rate as the melt zone passes through a transverse crack in the electrode. As the melt zone approaches the crack, poor heat conduction across the crack drives the local temperature in the electrode tip above its steady-state value, causing the controller to cut back on melting current in response to an increase in melting efficiency. The difficulty arises when the melt zone passes through the crack and encounters the relatively cold metal on the other side, giving rise to an abrupt drop in melt rate. This extremely dynamic melting situation is very difficult to handle using standard load-cell based melt rate control, resulting in large melt rate excursions. We have designed and tested a new generation melt rate controller that is capable of controlling melt rate through crack events. The controller is designed around an accurate dynamic melting model that uses four process variables: electrode tip thermal boundary layer, electrode gap, electrode mass and melting efficiency. Tests, jointly sponsored by the Specialty Metals Processing Consortium and Sandia National Laboratories, were performed at Carpenter Technology Corporation wherein two 0.43 m diameter Pyromet® 718 electrodes were melted into 0.51 m diameter ingots. Each electrode was cut approximately halfway through its diameter with an abrasive saw to simulate an electrode crack. Relatively accurate melt rate control through the cuts was demonstrated despite the observation of severe arc disturbances and loss of electrode gap control. Subsequent to remelting, one ingot was sectioned in the "as cast" condition, whereas the other was forged to 0.20 m diameter billet. Macrostructural characterization showed solidification white spots in regions affected by the cut in the electrode.

A new controller has been designed for vacuum arc remelting titanium alloys based on an accurate, low order, nonlinear, melting model. The controller adjusts melting current and electrode drive speed to match estimated gap and melt rate with operator supplied reference values. Estimates of gap and melt rate are obtained by optimally combining predictions from the model with measurements of voltage, current, and electrode position. Controller tests were carried out at Timet Corporation's Henderson Technical Laboratory in Henderson, Nevada. Previous test results were used to correlate measured gap to voltage and current. A controller test melt was performed wherein a 0.279 m diameter Ti-6Al-4V electrode was melted into 0.356 m diameter ingot. Commanded melt rate was varied from 20 to 90 g/s and commanded gap was held at 1.5 cm. Because no measure of electrode weight was available on the test furnace, electrode position data were analyzed and the results used to determine the actual melt rate. A gap-voltage-current factor space model was used to check estimated gap. The controller performed well, and both melt rate and electrode gap control were successfully demonstrated.

A variety of multivariate calibration algorithms for quantitative spectral analyses were investigated and compared, and new algorithms were developed in the course of this Laboratory Directed Research and Development project. We were able to demonstrate the ability of the hybrid classical least squares/partial least squares (CLSIPLS) calibration algorithms to maintain calibrations in the presence of spectrometer drift and to transfer calibrations between spectrometers from the same or different manufacturers. These methods were found to be as good or better in prediction ability as the commonly used partial least squares (PLS) method. We also present the theory for an entirely new class of algorithms labeled augmented classical least squares (ACLS) methods. New factor selection methods are developed and described for the ACLS algorithms. These factor selection methods are demonstrated using near-infrared spectra collected from a system of dilute aqueous solutions. The ACLS algorithm is also shown to provide improved ease of use and better prediction ability than PLS when transferring calibrations between near-infrared calibrations from the same manufacturer. Finally, simulations incorporating either ideal or realistic errors in the spectra were used to compare the prediction abilities of the new ACLS algorithm with that of PLS. We found that in the presence of realistic errors with non-uniform spectral error variance across spectral channels or with spectral errors correlated between frequency channels, ACLS methods generally out-performed the more commonly used PLS method. These results demonstrate the need for realistic error structure in simulations when the prediction abilities of various algorithms are compared. The combination of equal or superior prediction ability and the ease of use of the ACLS algorithms make the new ACLS methods the preferred algorithms to use for multivariate spectral calibrations.