ALBERT

Description: The use of hyperspectral data to determine the abundance of constituents in a certain portion of the Earth's surface relies on the capability of imaging spectrometers to provide a large amount of information at each pixel of a certain scene. Today, hyperspectral imaging sensors are capable of generating unprecedented volumes of radiometric data. The Airborne Visible/Infrared Imaging Spectrometer (AVIRIS), for example, routinely produces image cubes with 224 spectral bands. This undoubtedly opens a wide range of new possibilities, but the analysis of such a massive amount of information is not an easy task. In fact, most of the existing algorithms devoted to analyzing multispectral images are not applicable in the hyperspectral domain, because of the size and high dimensionality of the images. The application of neural networks to perform unsupervised classification of hyperspectral data has been tested by several authors and also by us in some previous work. We have also focused on analyzing the intrinsic capability of neural networks to parallelize the whole hyperspectral unmixing process. The results shown in this work indicate that neural network models are able to find clusters of closely related hyperspectral signatures, and thus can be used as a powerful tool to achieve the desired classification. The present work discusses the possibility of using a Self Organizing neural network to perform unsupervised classification of hyperspectral images. In sections 3 and 4, the topology of the proposed neural network and the training algorithm are respectively described. Section 5 provides the results we have obtained after applying the proposed methodology to real hyperspectral data, described in section 2. Different parameters in the learning stage have been modified in order to obtain a detailed description of their influence on the final results. Finally, in section 6 we provide the conclusions at which we have arrived.

Description: During the last several years, a number of airborne and satellite hyperspectral sensors have been developed or improved for remote sensing applications. Imaging spectrometry allows the detection of materials, objects and regions in a particular scene with a high degree of accuracy. Hyperspectral data typically consist of hundreds of thousands of spectra, so the analysis of this information is a key issue. Mathematical morphology theory is a widely used nonlinear technique for image analysis and pattern recognition. Although it is especially well suited to segment binary or grayscale images with irregular and complex shapes, its application in the classification/segmentation of multispectral or hyperspectral images has been quite rare. In this paper, we discuss a new completely automated methodology to find endmembers in the hyperspectral data cube using mathematical morphology. The extension of classic morphology to the hyperspectral domain allows us to integrate spectral and spatial information in the analysis process. In Section 3, some basic concepts about mathematical morphology and the technical details of our algorithm are provided. In Section 4, the accuracy of the proposed method is tested by its application to real hyperspectral data obtained from the Airborne Visible/Infrared Imaging Spectrometer (AVIRIS) imaging spectrometer. Some details about these data and reference results, obtained by well-known endmember extraction techniques, are provided in Section 2. Finally, in Section 5 we expose the main conclusions at which we have arrived.

Description: An approximate solution for the unsteady loading near the square-shape tip of a wing passing through an oblique gust is obtained in closed form. The aerodynamic theory developed can be used to predict airloads felt by a helicopter blade experiencing a blade/vortex interaction for high blade tip speed and/or for small vertical blade/vortex separation. Under these conditions one can show that the blade's trailing edge has little influence on the character of the chordwise loading at all spanwise sections; thus, the chord may be allowed to extend to infinity in the downstream direction. Therefore, the model considered here is that of a quarter-infinite flat plate wing with side edge passing subsonically through an oblique gust.

Description: A linear aerodynamic-acoustic theory is developed for the prediction of the surface pressure distribution and three-dimensional acoustic far-field for a flat plate rectangular wing encountering a stationary short-wavelength oblique gust. It is suggested that for an infinite-span wing, leading- and trailing-edge responses to a short-wavelength gust are essentially independent. This idea is used to solve for the two-dimensional pressure field due to the passage of an infinite-span wing through an oblique gust. By allowing the field point to come down to the wing's surface, one finds an expression for the surface pressure distribution which agrees with that given in the two-dimensional aerodynamic theories of Amiet and Adamczyk. Spanwise Fourier superposition of two-dimensional solutions to the infinite-span wing problem is used to approximate the three-dimensional acoustic field due to the interaction of a stationary oblique gust with a flat-plate rectangular wing traveling at a subsonic speed.

Description: Data from the altimeter onboard the European Remote Sensing Satellite (ERS1) was used to study the circulation of the Alboran Sea between -6 to 0 degrees and 35 to 38 degrees North. The results indicate that combining sea surface temperature data and sea level data from altimetry hold promise for understanding the circulation of the Western Mediterranean.

Description: The GeoEye Constellation consists of: a) IKONOS and OrbView-3 for high resolution; b) GeoEye with higher resolution 1Q2007; c) RESOUCESAT-1 for global crop assessment; d) OrbView-2 for ocean research and fish. IKONOS performance in 2005 included stable image quality, radiometry and geometric accuracy. reliability is 80% to 2008. Demonstrated capacity for high-volume, quick-response collection and production.

Description: A wind tunnel experiment involving single, double, and triple combinations of mutually interfering generic, unfinned aircraft stores has been conducted. Each combination of stores was tested at Mach numbers from 0.60 to 1.20 and at angles of attack from 0 to 25 deg for the single store and from 0 to 6 deg for the double and triple store configurations. Extensive axial and circumferential pressure and flow visualization data at each store location were obtained. Euler solutions for each configuration at 0 deg incidence have been generated and compared with experimental data. This comparison indicates an Euler flow solver can yield accurate predictions of the location and magnitude of multibody interference provided an appropriate grid is used and the viscous effects associated with these configurations remain small. The data indicate multibody interference in the transonic region increases as the freestream Mach number approaches 1 from either direction, and subsides as the Mach number moves away from sonic conditions. This interference is characterized by a large, localized reduction in pressure on the inboard surfaces of the bodies which results in forces that draw the configuration closer together.

Description: Computers have become essential tools for scientists simulating and observing nature. Simulations are formulated as mathematical models but are implemented as computer algorithms to simulate complex events. Observations are also analyzed and understood in terms of mathematical models, but the number of these observations usually dictates that we automate analyses with computer algorithms. In spite of their essential role, computers are also barriers to scientific understanding. Unlike hand calculations, automated computations are invisible and, because of the enormous numbers of individual operations in automated computations, the relation between an algorithm's input and output is often not intuitive. This problem is illustrated by the behavior of meteorologists responsible for forecasting weather. Even in this age of computers, many meteorologists manually plot weather observations on maps, then draw isolines of temperature, pressure, and other fields by hand (special pads of maps are printed for just this purpose). Similarly, radiologists use computers to collect medical data but are notoriously reluctant to apply image-processing algorithms to that data. To these scientists with life-and-death responsibilities, computer algorithms are black boxes that increase rather than reduce risk. The barrier between scientists and their computations can be bridged by techniques that make the internal workings of algorithms visible and that allow scientists to experiment with their computations. Here we describe two interactive systems developed at the University of Wisconsin-Madison Space Science and Engineering Center (SSEC) that provide these capabilities to Earth and space scientists.

Description: Characterising radiation from wildland fires is an important focus of fire science because radiation relates directly to the combustion process and can be measured across a wide range of spatial extents and resolutions. As part of a more comprehensive set of measurements collected during the 2012 Prescribed Fire Combustion and Atmospheric Dynamics Research (RxCADRE) field campaign, we used ground, airborne and spaceborne sensors to measure fire radiative power (FRP) from whole fires, applying different methods to small (2 ha) and large (.100 ha) burn blocks. For small blocks (n1/46), FRP estimated from an obliquely oriented long-wave infrared (LWIR) camera mounted on a boom lift were compared with FRP derived from combined data from tower-mounted radiometers and remotely piloted aircraft systems (RPAS). For large burn blocks (n1/43), satellite FRP measurements from the Moderate-resolution Imaging Spectroradiometer (MODIS) and Visible Infrared Imaging Radiometer Suite (VIIRS) sensors were compared with near-coincident FRP measurements derived from a LWIR imaging system aboard a piloted aircraft. We describe measurements and consider their strengths and weaknesses. Until quantitative sensors exist for small RPAS, their use in fire research will remain limited. For oblique, airborne and satellite sensors, further FRP measurement development is needed along with greater replication of coincident measurements, which we show to be feasible.