Real Time Intelligent Target Detection and Analysis with Machine Vision
We present an algoritm for detecting a specified set of targets for an Automatic Target Recognition (ATR) application.
Engineering topics
Publications and source records attributed to Brown, K..
We present an algoritm for detecting a specified set of targets for an Automatic Target Recognition (ATR) application.
Explore the source record for details and available documents.
Many applications that use hyperspectral imagery focus on dectection and recognition of targets that occupy a portion of a hyperspectral pixel.
Leaf Area Index (LAI) is one of the most commonly employed biophysical parameters used to characterize vegetation canopies and scale leaf physiological processes to larger scales. For example, LAI is a critical parameter used in regional scale estimates of evapotranspiration, photosynthesis, primary productivity, and carbon cycling (Running et al., 1989; Dorman and Sellers, 1989; Potter et al., 1993). LAI is typically estimated using ratio-based techniques, such as the Normalized Difference Vegetation Index (NDVI: e.g. Tucker 1979; Asrar et al., 1989; Sellers 1985, 1987). The physical basis behind this relationship depends on the high spectral contrast between scattered near-infrared (NIR) and absorbed red radiation in canopies. As the number of leaves present in a canopy increases over a unit area, NIR reflectance increases, while red reflectance decreases, resulting in an increase in the ratio. Through time series and image compositing, NDVI provides an additional temporal measure of how these parameters change, providing a means to monitor fluxes and productivity (Tucker et al., 1983). NDVI, while highly successful for agriculture and grassland ecosystems has been found to be less successful in evergreen chaparral and forested ecosystems (Badhwar et al., 1986; Gamon et al., 1993; Hall et al., 1995). Typically, the relationship between NDVI and LAI becomes progressively more asymptotic at LAI values above three (Sellers, 1985), although linear relationships have been observed in conifers at LAis as high as 13 (Spanner et al., 1990). In this paper, we explore an alternative approach for estimating LAI for remotely sensed data from AVIRIS based on estimates of canopy liquid water. Our primary objective is to test the hypothesis that the depth of the liquid water bands expressed in canopy reflectance spectra at 960, 1200, 1400 and 1900 nm increases with increasing LAI in canopies. This study builds from work by Roberts et al. (1997), in which liquid water was shown to increase following a gradient of increasing LAI ranging from grasslands to coniferous forests. In that study, it was observed that forests, which showed little variation in NDVI, showed significant variation in liquid water. In order to test this hypothesis, we analyzed field spectra measured over Populus resprouts of known LAI and monitored changes in liquid water in young Populus stands as they aged over a 4-year time span. The study was conducted in south-central Washington, in a clonal Populus fiber farm owned and operated by Boise-Cascade near the town of Wallula.
The goal of many applications that using hyperspectral images is to describe the constituent components of each pixel in a scene.
The VIGILANTE project is a planned vision system capable of tracking and recognizing targets in real time, on a small airborne platform.
In this paper, we explore an alternative approach for estimating LAI for remotely sensed data from AVIRIS based on estimates of canopy liquid water.
Microgravity has advantages for the cultivation of tissues with high fidelity; however, tissue formation requires cellular recognition and adhesion. We tested the hypothesis that simulated microgravity does not affect cell adhesion. Human colorectal carcinoma cells were cultured in the NASA Rotating Wall Vessel (RWV) under low shear stress with randomization of the gravity vector that simulates microgravity. After 6 - 7 days, cells were assayed for binding to various substrates and compared to cells grown in standard tissue culture flasks and static suspension cultures. The RWV cultures bound as well to basement membrane proteins and to Carcinoembryonic Antigen (CEA), an intercellular adhesion molecule, as control cultures did. Thus, microgravity does not alter epithelial cell adhesion and may be useful for tissue engineering.
A radiative transfer model for the bidirectional reflectance-distribution function (BRDF) shows that snow is moderately anisotropic in the near-infrared wavelengths. Although the directional-hemispherical albedo of snow decreases as the grains become larger, the forward scattering also increases, with the result that the illumination and viewing geometry must be considered when interpreting physical properties of the surface layer of the snow pack from remote sensing data. Measurements of the BRDF for a variety of snow conditions were made throughout the winter and spring seasons with a spectroradiometer, for wavelengths from 0.38 to 1.11 microns. Coincident with these the surface grain properties were analyzed by stereological methods. The data show that Mie scattering calculations using an equivalent spherical radius match the directional-hemispherical reflectance, but the BRDF usually shows a small peak in the backscattered direction that would not occur from spherical grains. The sphere with the same surface-to-volume ratio as the ice grains is used as the equivalent sphere.
A fully coupled, shock capturing technique is presented for chemically reacting flows at high Mach numbers. The technique makes use of a total variation diminishing (TVD) dissipation operator which results in sharp, crisp shocks. The eigenvalues and eigenvectors of the fully coupled system, which includes species conversion equations in addition to the gas dynamics equations, are analytically derived for a general reacting gas. Species production terms for a model dissociating gas are introduced and are included in the algorithm. The convective terms are solved using a first-order TVD scheme while the source terms are solved using a fourth-order Runge-Kutta scheme to enhance stability. Results from one-dimensional numerical experiments are shown for a two species and a three species gas.